JP6823612B2 - Predictors, prediction methods, and prediction programs - Google Patents

Description

The present invention relates to a prediction device, a prediction method, and a prediction program.

Conventionally, various techniques for identifying and predicting the appropriateness of a subject (behavior) who performs an action such as a user have been provided. For example, a task is provided to a target person (user), the characteristics of the target person are evaluated based on the performance value indicated by the target person in executing the task, and the target person is hired by an entity based on the evaluated characteristics. Techniques are provided to identify suitability for.

Special Table 2017-522676 Gazette

However, with the above-mentioned prior art, it may be difficult to appropriately predict the promising degree of the action subject. For example, it is difficult to predict the promising degree of the future potential of the user only by evaluating the user at the time of executing the task only by the evaluation based on the execution of the task provided to the target person (user).

The present application has been made in view of the above, and an object of the present application is to provide a prediction device, a prediction method, and a prediction program for appropriately predicting the promising degree of an action subject.

The prediction device according to the present application has an acquisition unit that acquires action information indicating a history of actions by a predetermined action subject, and the predetermined action device at a time point after the action based on the action information acquired by the acquisition unit. It is characterized by having a prediction unit that predicts the promising degree of the future of the action subject.

According to one aspect of the embodiment, there is an effect that the promising degree of the action subject can be appropriately predicted.

FIG. 1 is a diagram showing an example of a generation process according to an embodiment. FIG. 2 is a diagram showing an example of prediction processing according to the embodiment. FIG. 3 is a diagram showing a configuration example of the prediction system according to the embodiment. FIG. 4 is a diagram showing a configuration example of the prediction device according to the embodiment. FIG. 5 is a diagram showing an example of the learning data storage unit according to the embodiment. FIG. 6 is a diagram showing an example of a model information storage unit according to the embodiment. FIG. 7 is a diagram showing an example of a user information storage unit according to the embodiment. FIG. 8 is a diagram showing an example of a promising information storage unit according to the embodiment. FIG. 9 is a diagram showing an example of the prediction information storage unit according to the embodiment. FIG. 10 is a flowchart showing an example of the generation process according to the embodiment. FIG. 11 is a flowchart showing an example of the prediction process according to the embodiment. FIG. 12 is a hardware configuration diagram showing an example of a computer that realizes the function of the prediction device.

Hereinafter, a prediction device, a prediction method, and a mode for carrying out the prediction program (hereinafter referred to as “the embodiment”) according to the present application will be described in detail with reference to the drawings. The prediction device, prediction method, and prediction program according to the present application are not limited by this embodiment. Further, in each of the following embodiments, the same parts are designated by the same reference numerals, and duplicate description is omitted.

(Embodiment)
[1. Generation process]
First, an example of the generation process according to the embodiment will be described with reference to FIG. FIG. 1 is a diagram showing an example of a generation process according to an embodiment. In the example of FIG. 1, a case where the user is an action subject is shown as an example. The subject of action is not limited to the user as long as it is the subject of action, and may be various actors such as a predetermined group such as a school or a corporation. FIG. 1 shows a case where the prediction device 100 generates a model based on data including user behavior information and correct answer information indicating a position in an organization to which the user belongs, such as a company. In the example of FIG. 1, the prediction device 100 calculates the history of each user's behavior up to the year X (for example, 2010) and the score corresponding to the position of the user after the year X to Y (for example, 5 years). The case where the model is generated based on the data including the correct answer information shown is shown. The prediction target is not limited to the job title of the user, and may be any information. For example, the prediction target may be an evaluation value (score) in the company indicating how active the user is in the company or the like, that is, how useful the human resource is. For example, the evaluation value in the company may be personnel evaluation, sales results, award results in the company, research results, number of treatises, number of patents, and the like. Specifically, the evaluation values within the company are the scores corresponding to each rank of personnel evaluation, sales results and research achievements, the score according to the number of awards in the company, the score according to the number of papers, and the number of patents. It may be a score based on evaluations within various companies, such as a corresponding score. In the following, data including user behavior information and correct answer information associated with the behavior information will also be referred to as "learning data". The user behavior information will be described later.

In addition, the prediction device 100 predicts the promising degree of a user based on the behavior information of the user by using the generated model. Further, the prediction device 100 provides a service based on the information indicating the promising degree of the predicted action subject, and this point will be described with reference to FIG.

[Configuration of prediction system]
First, prior to the description of FIG. 1, the prediction system 1 shown in FIG. 3 will be described. As shown in FIG. 3, the prediction system 1 includes a terminal device 10 and a prediction device 100. The terminal device 10 and the prediction device 100 are communicably connected by wire or wirelessly via a predetermined network N. FIG. 3 is a diagram showing a configuration example of the prediction system according to the embodiment. The prediction system 1 shown in FIG. 3 may include a plurality of terminal devices 10 and a plurality of prediction devices 100.

The terminal device 10 is an information processing device used by the user. The terminal device 10 is realized by, for example, a smartphone, a tablet terminal, a notebook PC (Personal Computer), a desktop PC, a mobile phone, a PDA (Personal Digital Assistant), or the like. In the example shown in FIG. 1, a case where the terminal device 10 is a smartphone used by a user is shown. In the following, the terminal device 10 may be referred to as a user. That is, in the following, the user can be read as the terminal device 10. For example, FIG. 2 shows a case where the terminal device 10 is a smartphone used by the user U1. As described above, when "user U * (* is an arbitrary numerical value)" is described, it means that the user is a user identified by the user ID "U *". For example, when "user U1" is described, the user is a user identified by the user ID "U1".

The terminal device 10 may collect user behavior information and transmit it to the prediction device 100. Further, the terminal device 10 has a function such as a GPS (Global Positioning System) sensor, and may detect and acquire the user's position information (sensor information). Further, the terminal device 10 may detect various sensor information not only by a GPS sensor or the like but also by various sensors. For example, the terminal device 10 may predict and acquire the position information of the user by using the position information of the base station communicating with the user or the radio wave of WiFi (registered trademark) (Wireless Fidelity). Further, the terminal device 10 has a function of an acceleration sensor, and may detect and acquire acceleration information (sensor information) in the movement of the user. Further, the terminal device 10 may have various functions such as a temperature sensor and a barometric pressure sensor, and may be able to detect and acquire environmental information such as temperature and barometric pressure in which the user is placed. Further, the terminal device 10 may have various functions such as a heart rate sensor and may be able to detect and acquire the biometric information of the user. For example, a user who uses the terminal device 10 may be able to acquire the user's own context information by the terminal device 10 by wearing a wearable device capable of communicating with the terminal device 10. For example, a user who uses the terminal device 10 wears a wristband-type wearable device capable of communicating with the terminal device 10, and the terminal device 10 acquires information on the user's own heartbeat (pulse) by the terminal device 10. It may be possible.

The prediction device 100 is an information processing device that predicts the promising degree of the action subject based on the action information of the user. In the example of FIG. 2, the prediction device 100 is a model used for predicting the promising degree of the action subject, and predicts the promising degree of the action subject based on the output of the model in which the user's action information is input. For example, the prediction device 100 acquires the user's behavior information from the terminal device 10 used by the user. Further, for example, the prediction device 100 may acquire the user's behavior information from another information processing device that collects the user's behavior information.

Further, the prediction device 100 provides a service based on information on the predicted promisingness of the action subject. For example, the prediction device 100 may provide information indicating the promising degree of the action subject to an organization (organization) such as a company to which the user belongs or a company to which the user wishes to belong. In the example of FIG. 2, the prediction device 100 provides information indicating the promising position of the user U1 to the company in which the user U1 whose job title is predicted is taking the recruitment test. The prediction device 100 may provide information indicating the promising degree of the action subject to the business entity to which the user belongs.

First, in the example shown in FIG. 1, the learning data LD1 used by the prediction device 100 to generate the model M1 will be described. In the example of FIG. 1, as shown in the position group POS, which of the four positions, A position, B position, C position, and D position, is described as an example. In the example of FIG. 1, the A position is the director who has the highest rank, the B position is the manager who has the next highest rank after the A position, and the C position is the next highest rank position after the B position. It shows the case where there is no position which is the position of the section manager who is the lowest rank and the position D is the lowest rank.

For example, the prediction device 100 associates a higher score with a higher rank position. For example, as shown in the rank list RL11 (see FIG. 2), the prediction device 100 associates the highest score "1" with the A position (director) and the score "0.6" with the B position (manager). Attach, the score "0.3" is associated with the C position (section chief), and the lowest score "0" is associated with the D position (none). For example, the prediction device 100 stores information in which the job title and the score are associated with each other as shown in the rank list RL11 in the storage unit 120. The association with the position and the score as described above is an example, and the prediction device 100 may use the association of the score with various positions and each position.

In the example of FIG. 1, the prediction device 100 acquires the action information of each user as shown in the action group AC. For example, the learning data LD1 includes the behavior information ADT101 collected for the user U101 and the data DT101 including the correct answer information RDT101 indicating the position of the user U101. In the example of FIG. 1, the position of the user U101 after the years X to Y is the A position (director), and in the correct answer information RDT101, the position of the user U101 is the A position (director). Information indicating that the score is "1" is included.

For example, the prediction device 100 acquires the action information ADT 101 from the terminal device 10 used by the user U101. For example, the prediction device 100 acquires the action information ADT 101 including the history of various actions such as search by the user U101, content browsing, and purchase. For example, the prediction device 100 uses the behavior information ADT101 of the user U101 in one year of the year X and the correct answer information RDT101 indicating the position of the user U101 after the years X to Y as learning data. For example, the prediction device 100 uses the behavior information ADT101 of the user U101 in one year of 2010, which is the year X, and the correct answer information RDT101 indicating the position of the user U101 in 2015, which is five years after 2010, as learning data.

As described above, when "data DT * (* is an arbitrary numerical value)" is described, it means that the data is the data identified by the data ID "DT *". For example, when described as "data DT1", the data is data identified by the data ID "DT1".

The user behavior information referred to here includes various information related to the user behavior. For example, as shown in the behavior information ADT101, the user U101 is a user whose history of various behaviors such as search, content browsing, and purchase of the user U1 up to the year X (for example, 2010) has been collected. Shown. The user's action information is not limited to the above, and may include a history of various actions.

For example, the learning data LD1 includes the behavior information ATD102 collected for the user U102 and the data DT102 including the correct answer information RDT102 corresponding to the user U102. In the example of FIG. 1, the job title of the user U102 after X to Y years is the D job title (none), and in the correct answer information RDT102, the job title of the user U102 is the D job title (none). Information indicating that the score is "0" is included.

For example, the prediction device 100 acquires the action information ADT 102 from the terminal device 10 used by the user U 102. For example, the prediction device 100 acquires the action information ADT 102 including the history of various actions such as search by the user U102, content browsing, and purchase. For example, the prediction device 100 uses the behavior information ADT102 of the user U102 in one year of the year X and the correct answer information RDT102 indicating the position of the user U102 after the year X to Y as learning data. For example, the prediction device 100 uses the behavior information ADT102 of the user U102 in the one year of 2010, which is the year X, and the correct answer information RDT102 indicating the position of the user U102 in 2015 five years after 2010 as learning data.

Further, the learning data LD1 includes the behavior information ATD103 collected for the user U103, the data DT103 including the correct answer information RDT103 corresponding to the user U103, the data DT104 including the behavior information ATD104 collected for the user U104, and the like. For example, the prediction device 100 acquires the behavior information ATD103 from the terminal device 10 used by the user U103. Further, for example, the prediction device 100 acquires the action information ATD 104 from the terminal device 10 used by the user U 104.

Although only the data DT101 to DT104 are shown in FIG. 1, it is assumed that the learning data LD1 contains a large number of data. Further, for example, the prediction device 100 acquires the job title of each user from the company to which each user belongs. For example, the users U101 to U104 and the like may be users belonging to the company CP11 to which information is provided in FIG. Further, for example, the users U101 to U104 and the like may be users belonging to a company other than the company CP11.

For example, the prediction device 100 may generate correct answer information RDT101 to RDT104 and the like based on the information indicating the job titles of the users U101 to U104. Further, for example, the prediction device 100 may acquire the correct answer information RDT101 to RDT104 of each user U101 to U104 from the company to which each user belongs, another information processing device, or the like. Further, the prediction device 100 may acquire the behavior information of each user from an external device other than the terminal device 10 used by the user.

From here, each process of the prediction apparatus 100 will be described. First, the prediction device 100 acquires the learning data LD1 (step S11). Specifically, the prediction device 100 acquires the learning data LD1 including the data DT101 to DT104 and the like. In the example of FIG. 1, the prediction device 100 may acquire the learning data LD1 as described above from a predetermined external information processing device.

Then, the prediction device 100 adds a combination of the correct answer information included in the learning data LD1 acquired in step S11 and the user's behavior information as learning data (step S12). Specifically, the prediction device 100 adds the learning data LD1 to the learning data storage unit 121.

Then, the prediction device 100 generates a model based on the training data (step S13). In the example of FIG. 1, the prediction device 100 generates the model M1 as shown in the model list ML11. As described above, when "model M * (* is an arbitrary numerical value)" is described, it means that the model is a model identified by the model ID "M *". For example, when described as "model M1", the model is a model identified by the model ID "M1". For example, the prediction device 100 stores the information of the model M1 as shown in the model list ML11 in the model information storage unit 122 (see FIG. 6).

As shown in the model list ML11 in FIG. 1, the model M1 is a model used for predicting the degree of promising when the action subject is a user. Further, the model M1 indicates that the prediction target is the promising degree "position (promising promotion)", that is, the model used for predicting the promising promotion of the user who is the action subject. Further, the model M1 indicates that the specific model data is "model data MDT1".

For example, the prediction device 100 generates a model by performing learning using the data DT101 to DT104 in the learning data storage unit 121 as learning data (teacher data). For example, the prediction device 100 generates a model using features related to user behavior information. For example, the prediction device 100 generates a model using the feature amount related to the behavior information corresponding to the search by the user. For example, the prediction device 100 generates a model using the features related to the query used by the user for the search. Further, for example, the prediction device 100 generates a model by using the feature amount related to the behavior information corresponding to the content browsing by the user. For example, the prediction device 100 generates a model using features related to the content viewed by the user. Further, for example, the prediction device 100 generates a model by using the feature amount related to the behavior information corresponding to the purchase by the user. For example, the prediction device 100 generates a model using features related to goods and services purchased by the user.

In this way, the prediction device 100 learns the features corresponding to the user's behavior and generates a model for predicting the promising degree of the user who is the action subject. In the example of FIG. 1, the prediction device 100 learns the characteristics of the user's behavior information and generates a model for predicting the promising degree of the behavior subject. The feature corresponding to the user learned by the prediction device 100 may be input to the prediction device 100 by a person such as an administrator of the prediction device 100, or may be automatically learned (extracted) by the prediction device 100. Good.

For example, the prediction device 100 generates a model M1 using the training data LD1. For example, in the prediction device 100, since the title of the user U101 included in the correct answer information RDT101 is the title A (director) and the score is "1", the user behavior information ADT101 included in the data DT101 is input to the model M1. When this is done, the learning process is performed so that the score output by the model M1 approaches "1". For example, the prediction device 100 performs learning processing so that when the behavior information ADT101 of the data DT101 is input to the model M1, the score output by the model M1 approaches "1".

Further, for example, in the prediction device 100, since the title of the user U102 included in the correct answer information RDT102 is the D title (none) and the score is "0", the user behavior information ADT102 included in the data DT102 in the model M1. Is input, the learning process is performed so that the score output by the model M1 approaches "0". For example, the prediction device 100 performs a learning process so that when the behavior information ADT102 of the data DT102 is input to the model M1, the score output by the model M1 approaches "0".

Further, for example, in the prediction device 100, since the position of the user U103 included in the correct answer information RDT103 is the C position (section chief) and the score is "0.3", the behavior of the user included in the data DT103 in the model M1. When the information ADT103 is input, the learning process is performed so that the score output by the model M1 approaches "0.3". For example, the prediction device 100 performs learning processing so that when the behavior information ADT103 of the data DT103 is input to the model M1, the score output by the model M1 approaches "0.3".

Further, for example, in the prediction device 100, since the title of the user U104 included in the correct answer information RDT104 is the B title (manager) and the score is "0.6", the behavior of the user included in the data DT104 in the model M1. When the information ADT104 is input, the learning process is performed so that the score output by the model M1 approaches "0.6". For example, the prediction device 100 performs a learning process so that when the behavior information ADT104 of the data DT104 is input to the model M1, the score output by the model M1 approaches "0.6".

The model learning method is not limited to the above-mentioned method, and any known technique can be applied. It should be noted that the generation of each model may be performed by appropriately using various conventional techniques related to machine learning. For example, the model may be generated by using a technique related to machine learning of supervised learning such as SVM (Support Vector Machine). Also, for example, model generation may be performed using techniques related to machine learning in unsupervised learning. For example, model generation may be performed using a technique of deep learning. For example, the model may be generated by appropriately using various deep learning techniques such as DNN (Deep Neural Network), RNN (Recurrent Neural Network), and CNN (Convolutional Neural Network). The description regarding the generation of the model is an example, and the model may be generated by a learning method appropriately selected according to the information that can be acquired. That is, if the prediction device 100 can learn the model M1 so as to output the score corresponding to the correct answer information when the user's behavior information included in the learning data is input, any method can be used for the model M1. May be generated.

Through the above processing, the prediction device 100 generates the model M1 corresponding to the promising degree regarding the job title. For example, by inputting the user's behavior information into the model M1, the prediction device 100 causes the model M1 to output a score indicating the promising degree of the future position (promotion) of the user corresponding to the input user's behavior information. , Predict the promising position of the user based on the score output by the model M1. In this way, the prediction device 100 predicts the future activity of the user based on the score output by the model M1. For example, the prediction device 100 predicts how active the user will be in a company or society based on the score output by the model M1.

As described above, the prediction device 100 makes it possible to appropriately predict the promising degree of a certain user's job title by learning using the learning data in which the user's behavior information and the correct answer information about the job title are associated with each other. A model can be generated. Therefore, by using the model generated as described above, the prediction device 100 accurately predicts, for example, the promising future of how much a user is active in an organization (company) and is promoted. Becomes possible.

[2. Prediction processing]
An example of the prediction process according to the embodiment will be described with reference to FIG. FIG. 2 is a diagram showing an example of prediction processing according to the embodiment. FIG. 2 shows a case where the prediction device 100 predicts the promising degree of the user U1 and provides the information based on the prediction when the behavior information of the user U1 is given.

First, the prediction device 100 acquires the behavior information ADT1 of the user U1 (step S21). In the example of FIG. 2, the prediction device 100 acquires the behavior information ADT1 of the user U1 from the terminal device 10 used by the user U1. For example, the prediction device 100 acquires behavior information ADT1 including a history of various behaviors such as search by user U1, content browsing, and purchase. For example, the prediction device 100 acquires the behavior information ADT1 of the user U1 up to the year Z. For example, the prediction device 100 acquires the behavior information ADT1 of the user U1 in one year of the year Z. For example, the prediction device 100 acquires the behavior information ADT1 of the user U1 in one year of 2017, which is the year Z.

Then, for example, the prediction device 100 predicts the promising position of the user U1 in 2022, which is five years after 2017, by using the behavior information ADT1 of the user U1 and the model M1 in one year of 2017. The prediction device 100 that has acquired the behavior information ADT1 inputs the behavior information ADT1 into the model. For example, the prediction device 100 inputs the behavior information ADT1 of the user U1 into the model M1.

In the example of FIG. 2, the prediction device 100 generates (calculates) a score indicating the promising degree of the prediction target “post” of the user U1 by the processing as shown in the processing group PS22-1. The prediction device 100 inputs the behavior information of the user U1 who is the action subject into the model M1 corresponding to the prediction target “position (promising promotion)” (step S22-1). Specifically, the prediction device 100 inputs the behavior information ADT1 into the model M1. The model M1 to which the action information ADT1 is input outputs a score (step S23-1). In the example of FIG. 2, the model M1 to which the action information ADT1 is input outputs a score “0.7” as shown in the score SC11-1.

Then, the prediction device 100 predicts the promising degree of the user U1 based on the score output by the model M1 (step S24). For example, the prediction device 100 generates prediction information PL11 based on the prediction of the promising degree of the user U1. In the example of FIG. 2, the prediction device 100 predicts the promising degree of the user U1 based on the score output by the model M1 and the information stored in the promising degree information storage unit 124. The prediction device 100 is promising for the user U1 based on information in which the score and the promising degree are associated with each other (may be referred to as “association information”) as shown in the promising degree information storage unit 124 in FIG. Predict the degree.

The promising information storage unit 124 shown in FIG. 2 includes items such as “score” and “promising”. "Score" indicates the value (range) of the score corresponding to the corresponding promising degree. "Promising degree" indicates the level of promising degree of the behavioral subject when the score of the behavioral subject is the corresponding score. In the example of FIG. 2, the case where "promising degree" is three levels of "low", "medium", and "high" corresponding to the range of three scores is shown. For example, the associative information in the promising information storage unit 124 shown in FIG. 2 indicates that the promising degree of the user who is the action subject is "low" when the score is less than 0.3. The association information in the promising information storage unit 124 shown in FIG. 2 indicates that the promising degree of the user is “medium” when the score of the user is 0.3 or more and less than 0.6. Further, the association information in the promising information storage unit 124 shown in FIG. 2 indicates that the promising degree of the user is "high" when the score of the user is 0.6 or more.

The prediction device 100 compares the score output by the model M1 with the score range of the associated information in the promising information storage unit 124 shown in FIG. 2, and the score output by the model M1 falls within any score range. Determine if it is included. In the example of FIG. 2, the prediction device 100 determines that the score output by the model M1 is "0.7" and corresponds to the score "0.6 or more" corresponding to the promising degree "high". Therefore, as shown in the prediction information PL11, the prediction device 100 predicts that the promising degree corresponding to the prediction target “position (promising degree of promotion)” of the user U1 is “high”.

Further, the prediction device 100 predicts the job title of the user U1 after Z to Y years by using the rank list RL11. The rank list RL11 shown in FIG. 2 includes information in which the highest score "1" is associated with the A position (director) and the score "0.6" is associated with the B position (manager). Further, in the rank list RL11 shown in FIG. 2, information (promotion) in which the C position (section chief) is associated with the score "0.3" and the D position (none) is associated with the lowest score "0". Information) is included. For example, the prediction device 100 acquires information in which the job title and the score are associated with each other as shown in the rank list RL11 from the storage unit 120.

Then, the prediction device 100 compares the score output by the model M1 with the score of the promotion information in the rank list RL11 shown in FIG. 2, and determines which score or higher the score output by the model M1 is. In the example of FIG. 2, in the prediction device 100, the score output by the model M1 is "0.7", the score of the B position (manager) is "0.6" or more, and the score of the A position (director) is "0.7". It is determined that it is less than 1 ”. Therefore, as shown in the prediction information PL11, the prediction device 100 predicts the position of the user U1 in the prediction target “position (promising promotion)” as “manager or higher”. For example, the prediction device 100 predicts the position of the user U1 after Z to Y years as "manager or higher". In this case, the prediction device 100 predicts that the user U1 will be promoted to a position of manager or higher in the years Z to Y. For example, the prediction device 100 predicts that the user U1 will be promoted to a position of manager or higher in 2022, five years after 2017.

After that, the prediction device 100 provides information based on the predicted promisingness of the action subject (step S25). In the example of FIG. 2, the prediction device 100 provides the prediction information PL11 of the user U1 to a company that desires information regarding the promising degree of the user U1. In the example of FIG. 2, the prediction device 100 provides the prediction information PL11 of the user U1 to the business terminal 20 used by the manager MA1 of the company CP11 in which the user U1 is undergoing the recruitment test. The manager MA1 may be a person in charge of human resources of the company CP11. FIG. 2 shows a case where the business terminal 20 used by the administrator MA1 of the company CP11 to which the information is provided is a notebook computer. The business terminal 20 may be an information processing device similar to the terminal device 10. The prediction device 100 transmits the prediction information PL11 of the user U1 to the business terminal 20 used by the administrator MA1. Then, the business operator terminal 20 that has received the prediction information PL11 from the prediction device 100 displays the prediction information PL11. As a result, the manager MA1 of the company CP11 confirms the promising degree of the position of the user U1.

The manager MA1 of the company CP11 may request the prediction device 100 for information on the promising degree of the position of the user U1. In this case, the prediction device 100 may transmit the prediction information PL 11 to the business terminal 20 in response to a request for information regarding the promising degree of the user U1 from the business terminal 20 used by the administrator MA1.

As described above, the prediction device 100 predicts the promising degree of the action subject based on the action information of the user. In the example of FIG. 2, the prediction device 100 inputs the behavior information ADT101 of the user U1 into the model M1 to output a score regarding the promising degree of the prediction target “post” corresponding to the model M1. Then, the prediction device 100 predicts that the higher the score output by the model, the higher (greater) the promisingness of the user regarding the prediction target corresponding to the model. Then, the prediction device 100 provides the generated prediction information PL11 to the business terminal 20 used by the manager MA1 of the company CP11 in which the user U1 is undergoing the recruitment test. In this way, the prediction device 100 predicts the promising degree of the action subject, and transmits information on the predicted promising degree of the action subject to the provider who has a relationship with the action subject. As a result, the provider who receives the information provided by the prediction device 100 can appropriately grasp the promising degree of the action subject.

(2-1. Action subject, prediction target)
In the examples of FIGS. 1 and 2, the job title in the organization to which the user belongs is shown as a promising target, that is, the promising promotion target of the user is predicted, but the action subject and the promising target are various targets. There may be. For example, the action entity may be a predetermined corporation or a predetermined organization.

For example, the prediction target of the degree of promising may be any prediction target as long as it is a target showing the future activity of the action subject. For example, the promising prediction target may be a target showing the future potential of the action subject such as social activity. For example, when the action subject is a user, the promising target may be the promising degree in which the user starts a business and plays an active role (success), that is, the success promising degree of starting a business. For example, when the action subject is a corporation such as a corporation, the promising target may be the promising degree in which the performance of the corporation increases, that is, the business promising degree or the like.

(2-2. Behavior information)
In the examples of FIGS. 1 and 2, for the sake of simplicity, the information corresponding to the search, the content browsing, and the purchase is described as an example of the user's behavior information, but the user's behavior information includes various information. May be included. For example, the prediction device 100 is not limited to the user's behavior on the Internet (web), and may use various behavior information of the user that can be acquired. For example, the prediction device 100 may predict the promising degree by using the user's behavior information including the user's behavior information in the real space.

For example, the prediction device 100 may predict the promising degree by using the user's behavior information including the posted information posted by the user in a predetermined service. For example, the prediction device 100 may predict the promising degree by using the user's behavior information including the posted information posted by the user on the SNS. For example, the prediction device 100 may predict the promising degree by using the user's behavior information including the posted information posted by the user on Twitter (registered trademark), Facebook (registered trademark), or the like. For example, the prediction device 100 may predict the promising degree by using the user's behavior information including the character information posted by the user on Twitter. For example, the prediction device 100 may predict the promising degree by using the user's behavior information including the information input by the user in the message service. For example, the prediction device 100 may predict the promising degree by using the user's behavior information including the information input by the user in a message service such as LINE (registered trademark). For example, the prediction device 100 may predict the promising degree by using the behavior information of the user including the information input in the e-mail. Further, for example, if the information is input by the user, the prediction device 100 is promising by using the user's behavior information including various information such as text information, audio information, image information, and moving image information (video). You may predict the degree.

(2-3. About the model)
The generation of the model is not limited to FIG. 1, and may be in various modes. In the example of FIG. 1, a case where a plurality of models corresponding to each promising degree are generated is shown, but the prediction device 100 may generate one model that outputs a plurality of scores corresponding to each prediction target.

Further, in the example of FIG. 1, search, content browsing, and purchase are described as an example of user behavior information, but the prediction device 100 generates a model in which user behavior information including various information is input information. May be good.

For example, the prediction device 100 may generate a model in which the user's behavior information including the posted information posted by the user in a predetermined service is used as input information. For example, the prediction device 100 may generate a model in which the user's behavior information including the posted information posted by the user in the social networking service (SNS) is used as input information. For example, the prediction device 100 may generate a model in which the user's behavior information including the posted information posted by the user in Twitter, Facebook, or the like is used as input information. For example, the prediction device 100 may generate a model in which the user's behavior information including the character information posted by the user on Twitter is used as input information. For example, the prediction device 100 may generate a model in which the user's behavior information including the information input by the user in the message service is used as the input information. For example, the prediction device 100 may generate a model in which the user's behavior information including the information input by the user in a message service such as LINE is used as the input information. For example, the prediction device 100 may generate a model in which the user's behavior information including the information input in the e-mail is used as the input information. Further, for example, if the information is input by the user, the prediction device 100 uses the user's behavior information including various information such as text information, audio information, image information, and moving image information (video) as input information. You may generate a model to do.

(2-4. Prediction method)
In the example of FIG. 2, the case of predicting using a model is shown, but the promising degree of the user may be predicted without using the model. For example, the prediction device 100 predicts the promising degree of a user by determining whether or not the user's behavior satisfies the condition by using a condition indicating that the user is promising (also referred to as “promising condition”). You may. That is, the prediction device 100 may predict the promising degree of the user by determining whether or not the user's behavior corresponds to the promising condition. For example, when the prediction device 100 determines that the behavior of a user corresponds to a promising condition, the prediction device 100 may predict the user as having a high degree of promising. Further, for example, when the prediction device 100 determines that the behavior of the user does not correspond to the promising condition, the prediction device 100 may predict the user as having a "low" promising degree.

For example, the prediction device 100 may predict the promising degree of the user as "high" when the user concentrates on activities related to schoolwork or work during a predetermined period (for example, one month). .. For example, the prediction device 100 determines the promising degree of the user when the ratio of the position information of the user corresponding to the position of the university in the last three months until graduation from the university is equal to or more than a predetermined threshold value (for example, 50%). It may be predicted to be "high". The above is an example, and the prediction device 100 may predict the promising degree of the user by determining whether the user's behavior corresponds to various promising conditions.

[3. Predictor configuration]
Next, the configuration of the prediction device 100 according to the embodiment will be described with reference to FIG. FIG. 4 is a diagram showing a configuration example of the prediction device according to the embodiment. As shown in FIG. 4, the prediction device 100 includes a communication unit 110, a storage unit 120, and a control unit 130. The prediction device 100 has an input unit (for example, a keyboard, a mouse, etc.) that receives various operations from the administrator of the prediction device 100, and a display unit (for example, a liquid crystal display, etc.) for displaying various information. You may.

(Communication unit 110)
The communication unit 110 is realized by, for example, a NIC (Network Interface Card) or the like. Then, the communication unit 110 is connected to the network by wire or wirelessly, and transmits / receives information to / from the terminal device 10.

(Memory unit 120)
The storage unit 120 is realized by, for example, a semiconductor memory element such as a RAM (Random Access Memory) or a flash memory (Flash Memory), or a storage device such as a hard disk or an optical disk. As shown in FIG. 4, the storage unit 120 according to the embodiment includes a learning data storage unit 121, a model information storage unit 122, a user information storage unit 123, a promising degree information storage unit 124, and a prediction information storage unit 125. And have.

(Learning data storage unit 121)
The learning data storage unit 121 according to the embodiment stores various information related to the learning data. FIG. 5 is a diagram showing an example of the learning data storage unit according to the embodiment. For example, the learning data storage unit 121 stores teacher data used for generating a model. The learning data storage unit 121 shown in FIG. 5 includes items such as "data ID", "user ID", "correct answer information", and "behavior information".

The "data ID" indicates identification information for identifying the data. For example, the "data ID" indicates identification information for identifying a combination of user behavior information and correct answer information used as learning data (teacher data). For example, the data identified by the data ID "DT101" corresponds to the data DT101 shown in the example of FIG. The "user ID" indicates identification information for identifying a user who is an action subject.

“Correct answer information” indicates correct answer information corresponding to the data identified by the data ID. For example, the "correct answer information" indicates the job title of the user corresponding to the data identified by the data ID and the score corresponding to the job title. In FIG. 5, an example in which conceptual information such as “RDT101” is stored in “correct answer information” is shown, but in reality, information indicating the user's job title and the score corresponding to that job title (for example, A job title or score “A” 1 ”etc.) Or a file path name indicating the storage location is stored.

The "behavior information" indicates the behavior information including the history of the user's behavior corresponding to the data identified by the data ID. In FIG. 5, an example in which conceptual information such as "ADT101" is stored in "behavior information" is shown, but in reality, various information about the behavior performed by the user or a file path name indicating the storage location thereof. Etc. are stored.

For example, in the "behavior information", the behavior information including the history of the behavior of the user who is the behavior subject on the Internet (web) is stored. For example, the "behavior information" stores behavior information including a search history indicating what kind of search query the user used to perform the search. For example, in the "behavior information", behavior information including a purchase history indicating what kind of product the user has purchased is stored. Further, for example, in the "behavior information", behavior information including a browsing history indicating what kind of content the user has browsed is stored. The above is an example, and the history of various actions of the action subject is stored in the "behavior information".

For example, in the example shown in FIG. 5, the data (data DT101) identified by the data ID "DT101" is the information (data) corresponding to the user (user U101) identified by the user ID "U101". Shown. For example, in the example shown in FIG. 5, it is shown that the data DT 101 includes the behavior information ADT 101 corresponding to the user U 101. Further, the data DT 101 indicates that the correct answer information RDT 101 is included.

The learning data storage unit 121 is not limited to the above, and may store various information depending on the purpose. For example, the learning data storage unit 121 may store information regarding the date and time when the learning data was added. Further, for example, when the user's attribute information is used for inputting the model, the learning data storage unit 121 may store the information regarding the user's attribute. For example, the learning data storage unit 121 may store demographic attribute information and psychographic attribute information as attribute information. For example, the learning data storage unit 121 may store information such as name, age, gender, home, place of work, hobbies, interests, family structure, income, friendships such as friends, and lifestyle.

In the example shown in FIG. 5, the learning data storage unit 121 stores the learning data used to generate the model M1 of the prediction target “position”, but the learning data storage unit 121 stores each prediction target. The learning data corresponding to is stored. For example, the learning data storage unit 121 may store learning data in which a plurality of correct answer information corresponding to each prediction target is associated with the behavior information of each user. For example, the behavior information ADT 101 of the user U101 may be associated with the correct answer information (correct answer information RDT201) of the prediction target "entrepreneur" in addition to the correct answer information RDT101 of the prediction target "post". For example, the prediction device 100 uses the behavior information ADT101 of the user U101 and the correct answer information RDT201 to generate a model M2 corresponding to the prediction target “entrepreneurship”. Further, in the example shown in FIG. 5, the case where the learning data storage unit 121 stores the user's behavior information as learning data is shown, but the learning data storage unit 121 is not limited to the user's behavior information, and various types such as corporations and the like. The behavioral information of the behavioral subject may be stored as learning data. For example, the prediction device 100 generates a model M3 corresponding to the prediction target "performance" by using the behavior information of the corporation and the correct answer information corresponding to the prediction target "performance".

(Model information storage unit 122)
The model information storage unit 122 according to the embodiment stores information about the model. For example, the model information storage unit 122 stores model information (model data) generated by the generation process. FIG. 6 is a diagram showing an example of a model information storage unit according to the embodiment. The model information storage unit 122 shown in FIG. 6 includes items such as "action subject", "model ID", "prediction target", and "model data". Although only the models M1 to M3 are shown in FIG. 6, a large amount of model information such as M4 and M5 may be stored according to each prediction target.

In addition, each model may be expected to be used as a program module that is a part of artificial intelligence software. For example, the model M1 may be expected to be used as a program module for predicting the prediction target “position”. For example, the model data MDT1 may be data used as a program module for predicting the prediction target “position”. For example, the model M1 causes the computer to function to output a quantified value of the user's promising degree based on the user's behavior information.

“Action subject” indicates a behavioral subject whose promising degree is predicted. The "model ID" indicates identification information for identifying the model. For example, the model identified by the model ID "M1" corresponds to the model M1 shown in the example of FIG. “Prediction target” indicates a target for which the corresponding model predicts the promising degree. Further, "model data" indicates data of the associated corresponding model. For example, "model data" includes information including nodes in each layer, functions adopted by each node, connection relationships of nodes, and connection coefficients set for connections between nodes.

For example, in the example shown in FIG. 6, in the model (model M1) identified by the model ID “M1” and the model (model M2) identified by the model ID “M2”, the target action subject is the user (natural person). Indicates that. Further, for example, in the example shown in FIG. 6, the model (model M3) identified by the model ID “M3” indicates that the target action subject is a corporation.

For example, in the example shown in FIG. 6, in the model M1, the prediction target is the “position (probability of promotion)”, and the degree of promising of the position of the user, that is, the degree of the user is based on the input behavior information of the user. It is shown that it is used to predict the promising degree (promising degree of promotion) of promotion. Further, it is shown that the model data of the model M1 is the model data MDT1.

The model M1 (model data MDT1) is any of an input layer, an output layer, and an input layer to an output layer into which the action information of the user (action subject) is input, and belongs to a layer other than the output layer. An output layer for the action information of the user (action subject) input to the input layer, including the first element and the second element whose value is calculated based on the weights of the first element and the first element. By performing an operation based on the weights of the first element and the first element with each element belonging to each layer other than the first element as the first element, the value of the score used for predicting the promising degree of the user (action subject) is output as the output layer. It is a model for making the computer function so that it outputs from.

Here, it is assumed that the models M1 to M3 and the like are realized by the regression model represented by "y = a1 * x1 + a2 * x2 + ... + ai * xi". In this case, for example, the first element included in the model M1 corresponds to input data (xi) such as x1 and x2. Further, the weight of the first element corresponds to the coefficient ai corresponding to xi. Here, the regression model can be regarded as a simple perceptron having an input layer and an output layer. When each model is regarded as a simple perceptron, the first element corresponds to any node of the input layer, and the second element can be regarded as the node of the output layer.

Further, it is assumed that the models M1 to M3 and the like are realized by a neural network having one or a plurality of intermediate layers such as DNN. In this case, for example, the first element included in the model M1 corresponds to either the node of the input layer or the intermediate layer. Further, the second element corresponds to a node in the next stage, which is a node to which a value is transmitted from a node corresponding to the first element. Further, the weight of the first element corresponds to a connection coefficient which is a weight considered for the value transmitted from the node corresponding to the first element to the node corresponding to the second element.

The model information storage unit 122 is not limited to the above, and may store various model information depending on the purpose.

(User information storage unit 123)
The user information storage unit 123 according to the embodiment stores various information about the user. FIG. 7 is a diagram showing an example of a user information storage unit according to the embodiment. For example, the user information storage unit 123 stores various behavior information of each user. The user information storage unit 123 shown in FIG. 7 includes items such as "user ID" and "behavior information".

The "user ID" indicates identification information for identifying a user. For example, the user identified by the user ID "U1" corresponds to the user U1 shown in the example of FIG. The "behavior information" indicates the behavior information which is the history of the user's behavior corresponding to the data identified by the data ID.

The "behavior information" indicates the behavior information including the history of the user's behavior corresponding to the data identified by the data ID. In FIG. 5, an example in which conceptual information such as "ADT1" is stored in "behavior information" is shown, but in reality, various information about the behavior performed by the user or a file path name indicating the storage location thereof. Etc. are stored.

For example, in the "behavior information", the behavior information including the history of the behavior of the user who is the behavior subject on the Internet (web) is stored. For example, the "behavior information" stores behavior information including a search history indicating what kind of search query the user used to perform the search. For example, in the "behavior information", behavior information including a purchase history indicating what kind of product the user has purchased is stored. Further, for example, in the "behavior information", behavior information including a browsing history indicating what kind of content the user has browsed is stored. The above is an example, and the history of various actions of the action subject is stored in the "behavior information".

For example, in the example shown in FIG. 7, the action information corresponding to the user (user U1) identified by the user ID "U1" includes a history of various actions such as search, content browsing, and purchase performed by the user U1. Indicates that the behavior information is ADT1. Further, for example, in the example shown in FIG. 7, the action information corresponding to the user (user U2) identified by the user ID "U2" is a history of various actions such as search, content browsing, and purchase performed by the user U2. Indicates that the behavior information ADT2 includes.

The user information storage unit 123 is not limited to the above, and may store various information depending on the purpose. Further, although FIG. 7 shows a case where the action information is stored in the user information storage unit 123 for each user ID, the action information is not limited to each user ID and may be stored, for example, in chronological order.

The user information storage unit 123 is not limited to the above, and may store various information depending on the purpose. For example, the user information storage unit 123 may store information (attribute information) related to the user's attributes as user information. For example, the user information storage unit 123 may store demographic attribute information and psychographic attribute information as attribute information. For example, the user information storage unit 123 may store information such as name, age, gender, home, work location, hobbies, interests, family structure, income, friendships such as friends, and lifestyle as attribute information. Good.

(Promisability information storage unit 124)
The promising information storage unit 124 according to the embodiment stores various information regarding the promising degree. FIG. 8 is a diagram showing an example of a promising information storage unit according to the embodiment. For example, the promising information storage unit 124 stores information indicating the correspondence between the score and the promising degree. The promising information storage unit 124 shown in FIG. 8 includes items such as “score” and “promising”.

"Score" indicates the value (range) of the score corresponding to the corresponding promising degree. "Promising degree" indicates the level of promising degree of the behavioral subject when the score of the behavioral subject is the corresponding score. In the example of FIG. 8, the case where the “promising degree” is three levels of “low”, “medium”, and “high” corresponding to the range of three scores is shown.

For example, in the example shown in FIG. 8, when the score of the action subject is less than 0.3, it indicates that the promising degree of the action subject is “low”. When the score of the actor is 0.3 or more and less than 0.6, it indicates that the promising degree of the actor is "medium". Moreover, when the score of the action subject is 0.6 or more, it indicates that the promising degree of the action subject is “high”.

The promising information storage unit 124 is not limited to the above, and various information may be stored depending on the purpose. For example, the promising information storage unit 124 may store information indicating the correspondence between the score and the promising degree for each prediction target.

(Prediction information storage unit 125)
The prediction information storage unit 125 according to the embodiment stores various information related to prediction. FIG. 9 is a diagram showing an example of the prediction information storage unit according to the embodiment. For example, the prediction information storage unit 125 stores various information regarding the promising degree of the predicted action subject. The prediction information storage unit 125 shown in FIG. 9 includes items such as "action subject", "ID", "prediction target", "score", "prediction rank", and "promising degree".

“Action subject” indicates a behavioral subject whose promising degree is predicted. The "ID" indicates identification information for identifying an action subject such as a user or a corporation. For example, the action subject (user) identified by the ID "U1" corresponds to the user U1 shown in the example of FIG. “Prediction target” indicates an target for which the degree of promising is predicted. "Score" indicates the score of the corresponding prediction target for the corresponding action subject. The "predicted rank" indicates the predicted rank of the corresponding action subject based on the score of the corresponding predicted target. The “promising degree” indicates the promising degree predicted based on the score of the corresponding predicted object for the corresponding action subject.

For example, in the example shown in FIG. 9, the user U1 who is a user (natural person) identified by the user ID "U1" has a score of "0.7" for the prediction target "position (promising promotion)". Is shown. The user U1 indicates that the prediction target "position (promising degree of promotion)" is a predicted rank "manager or higher", that is, a user whose job title is predicted to be promoted to manager or higher. Further, the user U1 indicates that the user has a high degree of promising for the prediction target “position (promising degree of promotion)”, that is, a user with a high degree of promising for the position.

The prediction information storage unit 125 is not limited to the above, and may store various information depending on the purpose.

(Control unit 130)
Returning to the description of FIG. 4, the control unit 130 is a controller, and is stored in a storage device inside the prediction device 100 by, for example, a CPU (Central Processing Unit) or an MPU (Micro Processing Unit). It is realized by executing various programs (corresponding to an example of a prediction program) using the RAM as a work area. Further, the control unit 130 is a controller, and is realized by, for example, an integrated circuit such as an ASIC (Application Specific Integrated Circuit) or an FPGA (Field Programmable Gate Array). The control unit 130 is either an input layer, an output layer, or an input layer to an output layer in which the action information of the action subject is input by information processing according to the model M1 or the like stored in the model information storage unit 122. The first element belonging to the layer other than the output layer and the second element whose value is calculated based on the weights of the first element and the first element are included and input to the input layer. For the action information, each element belonging to each layer other than the output layer is set as the first element, and the score used for predicting the promising degree of the action subject is calculated by performing the calculation based on the weights of the first element and the first element. Output the value from the output layer.

As shown in FIG. 4, the control unit 130 includes an acquisition unit 131, a generation unit 132, a prediction unit 133, and a provision unit 134, and realizes or executes an information processing function or operation described below. .. The internal configuration of the control unit 130 is not limited to the configuration shown in FIG. 4, and may be another configuration as long as it performs information processing described later. The control unit 130 includes a second element whose value is calculated based on the first element and the weight of the first element by information processing according to the model M1 (model data MDT1) stored in the storage unit 120. By performing an operation based on the weights of the first element and the first element with each element belonging to each layer other than the output layer as the first element for the user's action information input to the input layer including The score value used for predicting the promising degree of the subject is output from the output layer.

(Acquisition unit 131)
The acquisition unit 131 acquires various types of information. For example, the acquisition unit 131 acquires various information from the learning data storage unit 121, the model information storage unit 122, the user information storage unit 123, the promising information storage unit 124, the prediction information storage unit 125, and the like. Further, the acquisition unit 131 may acquire various information from an external information processing device. Further, the acquisition unit 131 may acquire various information from the terminal device 10 or the like. For example, the acquisition unit 131 acquires the user's behavior information from the terminal device 10 used by the user. Further, for example, the acquisition unit 131 may acquire the user's behavior information from another information processing device that collects the user's behavior information.

The acquisition unit 131 acquires action information indicating a history of actions by a predetermined action subject. The acquisition unit 131 acquires a model that outputs a score indicating the promising degree of the predetermined action subject in response to the input of the action information of the predetermined action subject. The acquisition unit 131 acquires the action information of a predetermined action subject who is a user. The acquisition unit 131 acquires behavior information including a history of user behavior on the Internet. The acquisition unit 131 acquires behavior information including a search history by the user.

The acquisition unit 131 acquires behavior information including a purchase history by the user. The acquisition unit 131 acquires behavioral information of actions by a predetermined action subject, which is a predetermined group. The acquisition unit 131 acquires behavioral information of actions by a predetermined action subject, which is a predetermined corporation.

Further, for example, the acquisition unit 131 acquires user information including user attribute information. For example, the acquisition unit 131 acquires user information including information indicating the gender, age, and the like of the user. For example, the acquisition unit 131 acquires user information including information indicating demographic attribute information and psychographic attribute information. For example, the acquisition unit 131 acquires user information including information indicating a name, age, gender, home, work location, hobbies, interests, family structure, income, friendships such as friends, lifestyle, and the like.

In the example of FIG. 1, the acquisition unit 131 acquires the action information of each user U101 to U104.
The acquisition unit 131 acquires the action information ADT 101 from the terminal device 10 used by the user U101. The acquisition unit 131 acquires the action information ADT 101 including the history of various actions such as search by the user U101, content browsing, and purchase. The acquisition unit 131 acquires the action information ADT 102 from the terminal device 10 used by the user U 102. The acquisition unit 131 acquires the action information ADT 102 including the history of various actions such as search by the user U102, content browsing, and purchase. The acquisition unit 131 acquires the action information ATD 103 from the terminal device 10 used by the user U 103. Further, the acquisition unit 131 acquires the action information ATD 104 from the terminal device 10 used by the user U 104.

The acquisition unit 131 acquires the job title of each user from the company to which each user belongs. The acquisition unit 131 may acquire the correct answer information RDT101 to RDT104 of each user U101 to U104 from the company to which each user belongs, another information processing device, or the like. Further, the prediction device 100 may acquire the behavior information of each user from an external device other than the terminal device 10 used by the user.

For example, the acquisition unit 131 acquires the action information ADT 101 indicating the history of the actions of the user U 101 from the terminal device 10 used by the user U 101. For example, the acquisition unit 131 acquires the action information ADT 102 indicating the action history of the user U 102 from the terminal device 10 used by the user U 102. For example, the acquisition unit 131 acquires the action information ADT 103 indicating the action history of the user U 103 from the terminal device 10 used by the user U 103. Further, for example, the acquisition unit 131 acquires the action information ADT 104 indicating the action history of the user U104 from the terminal device 10 used by the user U104.

In the example of FIG. 1, the acquisition unit 131 acquires the learning data LD1. For example, the acquisition unit 131 acquires the learning data LD1 including the data DT101 to DT104.

In the example of FIG. 2, the acquisition unit 131 acquires the action information ADT1 of the user U1. The acquisition unit 131 acquires the action information ADT1 of the user U1 from the terminal device 10 used by the user U1. The acquisition unit 131 acquires the action information ADT1 including the history of various actions such as search by the user U1, content browsing, and purchase. The acquisition unit 131 acquires the behavior information ADT1 of the user U1 up to the year Z. The acquisition unit 131 acquires the behavior information ADT1 of the user U1 in one year of the year Z. The acquisition unit 131 acquires the behavior information ADT1 of the user U1 in one year of 2017, which is the year Z. For example, the acquisition unit 131 compares the date and time (acquisition date and time) when each behavior information is acquired (detected) with the target period (for example, 2017) from the user's behavior information group, and the acquisition date and time is the target. If it is included in the period, the action information is acquired as the action information to be acquired. The acquisition unit 131 acquires information in which the job title and the score are associated with each other as shown in the rank list RL11 from the storage unit 120.

(Generator 132)
The generation unit 132 generates various information. For example, the generation unit 132 uses the learning data stored in the learning data storage unit 121 to generate a model as shown in the model information storage unit 122. For example, the generation unit 132 generates a model used for predicting the promising degree of the action subject based on the learning data acquired by the acquisition unit 131. For example, the generation unit 132 generates a model used for predicting the promising degree of the action subject based on the learning data including the user's behavior information and the correct answer information indicating the type of promising degree associated with the user. ..

For example, the generation unit 132 generates models M1 to M3 and the like, and stores the generated models M1 to M3 and the like in the model information storage unit 122. The generation unit 132 may generate models M1 to M3 and the like using any learning algorithm. For example, the generation unit 132 generates models M1 to M3 and the like by using learning algorithms such as a neural network, a support vector machine (SVM), clustering, and reinforcement learning. As an example, when the generation unit 132 generates models M1 to M3 and the like using a neural network, the models M1 to M3 and the like include an input layer containing one or more neurons and an intermediate layer containing one or more neurons. It has an output layer containing the above neurons.

The generation unit 132 generates a model and stores the generated model in the model information storage unit 122. Specifically, the generation unit 132 is a first layer that is any of an input layer, an output layer, and an input layer to an output layer into which the action information of the action subject is input, and belongs to a layer other than the output layer. Each element belonging to each layer other than the output layer is included with respect to the action information input to the input layer, including the element and the second element whose value is calculated based on the first element and the weight of the first element. As the first element, a model is generated in which the score value used for predicting the promising degree of the action subject is output from the output layer by performing the calculation based on the first element and the weight of the first element.

For example, the generation unit 132 generates a model based on the training data. For example, the generation unit 132 generates models M1 to M3 and the like as shown in the model information storage unit 122. In the example of FIG. 1, the generation unit 132 generates the model M1 based on the training data. For example, the generation unit 132 generates a model by a predetermined learning process using the learning data. For example, the generation unit 132 generates a model by performing learning using the data DT101 to DT104 in the learning data storage unit 121 as learning data (teacher data).

In the example of FIG. 1, the generation unit 132 generates a model using the feature amount related to the user's behavior information. The generation unit 132 generates a model using the feature amount related to the behavior information corresponding to the search by the user. The generation unit 132 generates a model using the features related to the query used by the user for the search. In addition, the generation unit 132 generates a model using the feature amount related to the behavior information corresponding to the content browsing by the user. The generation unit 132 generates a model using the feature amount related to the content viewed by the user. In addition, the generation unit 132 generates a model using the feature amount related to the behavior information corresponding to the purchase by the user. The generation unit 132 generates a model using the features related to the goods and services purchased by the user.

In the example of FIG. 1, the generation unit 132 learns the features corresponding to the user's behavior and generates a model for predicting the promising degree of the user who is the action subject. The generation unit 132 learns the characteristics of the user's behavior information and generates a model for predicting the promising degree of the behavior subject.

The generation unit 132 generates the model M1 using the learning data LD1. In the generation unit 132, the title of the user U101 included in the correct answer information RDT101 is the title A (director) and the score is "1". Therefore, the user behavior information ADT101 included in the data DT101 is input to the model M1. In this case, the learning process is performed so that the score output by the model M1 approaches "1". The generation unit 132 performs learning processing so that when the behavior information ADT101 of the data DT101 is input to the model M1, the score output by the model M1 approaches "1".

Further, in the generation unit 132, since the title of the user U102 included in the correct answer information RDT102 is the D title (none) and the score is "0", the user behavior information ADT102 included in the data DT102 is input to the model M1. If so, the learning process is performed so that the score output by the model M1 approaches "0". The generation unit 132 performs learning processing so that when the behavior information ADT102 of the data DT102 is input to the model M1, the score output by the model M1 approaches "0".

Further, in the generation unit 132, since the position of the user U103 included in the correct answer information RDT103 is the C position (section chief) and the score is "0.3", the user behavior information ADT103 included in the data DT103 in the model M1 Is input, the learning process is performed so that the score output by the model M1 approaches "0.3". The generation unit 132 performs learning processing so that when the action information ADT103 of the data DT103 is input to the model M1, the score output by the model M1 approaches "0.3".

Further, in the generation unit 132, since the position of the user U104 included in the correct answer information RDT104 is the B position (manager) and the score is "0.6", the user behavior information ADT104 included in the data DT104 in the model M1 Is input, the learning process is performed so that the score output by the model M1 approaches "0.6". The generation unit 132 performs the learning process so that the score output by the model M1 approaches "0.6" when the action information ADT104 of the data DT104 is input to the model M1.

In the example of FIG. 2, the generation unit 132 generates (calculates) a score indicating the promising degree of the prediction target “post” of the user U1 by the processing as shown in the processing group PS22-1. The generation unit 132 generates the prediction information PL11 based on the prediction of the promising degree of the user U1.

(Prediction unit 133)
The prediction unit 133 predicts various types of information. The prediction unit 133 predicts various information using the information stored in the storage unit 120. The prediction unit 133 uses various information stored in the learning data storage unit 121, the model information storage unit 122, the user information storage unit 123, the promising information storage unit 124, the prediction information storage unit 125, and the like. Predict information. For example, the prediction unit 133 predicts various information based on various information acquired by the acquisition unit 131. The prediction unit 133 inputs the action information (input data) of the action subject such as a user or a corporation into the input layer of the model by information processing according to the model such as the model M1. Then, the prediction unit 133 propagates the input data to the intermediate layer and the output layer, so that the output layer outputs a score indicating the promising degree of the action subject.

The prediction unit 133 predicts the promising degree of the future potential of a predetermined action subject at a time point after the action based on the action information acquired by the acquisition unit 131. The prediction unit 133 inputs the behavior information into the model, and predicts the promising degree of a predetermined behavior subject based on the score output by the model. The prediction unit 133 predicts the promising degree of a user who is a human (natural person) by using the behavior information of a predetermined action subject who is the user.

The prediction unit 133 appropriately predicts the promising degree of the user by using the behavior information including the history of the user's behavior on the Internet. The prediction unit 133 predicts the promising degree of the user by using the behavior information including the search history by the user. The prediction unit 133 predicts the promising degree of the user by using the behavior information including the purchase history by the user.

The prediction unit 133 predicts the promising degree of the user's job title. The prediction unit 133 predicts the promising degree of the user's job title and predicts whether the user will be promoted in an organization such as a company. The prediction unit 133 predicts the promising degree of the user by comparing and determining which importance of the correspondence table the score output by the model corresponds to by using the correspondence table between the score and the promising degree.

The prediction unit 133 predicts the promising degree of a predetermined group by using the action information of the action by the predetermined action subject which is a predetermined group. The prediction unit 133 predicts the promising degree of a predetermined organization such as an organization composed of a plurality of user groups and the like. The prediction unit 133 predicts the promising degree of a predetermined corporation by using the behavior information of the behavior by the predetermined action subject which is the predetermined corporation. The prediction unit 133 predicts the promising degree of a predetermined corporation, which is an action subject having legal personality.

Further, for example, the prediction unit 133 may predict the promising degree of the user by using the user information including the attribute information of the user. For example, the prediction unit 133 may predict the promising degree of the user by using the user information including the information indicating the gender, age, and the like of the user. For example, the prediction unit 133 may predict the promising degree of the user by using the demographic attribute information and the psychographic attribute information. For example, the prediction unit 133 predicts the promising degree of a user by using information indicating name, age, gender, home, place of work, hobbies, interests, family structure, income, friendships such as friends, lifestyle, etc. You may.

For example, the prediction unit 133 predicts the promising position of the user U1 in 2022, which is five years after 2017, by using the behavior information ADT1 of the user U1 and the model M1 in the one year of 2017. The prediction unit 133 inputs the behavior information ADT1 into the model M1. For example, the prediction unit 133 compares the date and time (acquisition date and time) when each of the action information is acquired (detected) with the target period (for example, 2017) from the action information group of the user U1, and targets the acquisition date and time. If it is included in the period of, the action information is extracted as the action information to be acquired. Then, the prediction unit 133 uses the extracted behavior information as information (behavior information ADT1) to be input to the model M1.

In the example of FIG. 2, the prediction unit 133 generates (calculates) a score indicating the promising degree of the prediction target “post” of the user U1 by the processing as shown in the processing group PS22-1. The prediction unit 133 inputs the behavior information of the user U1 who is the action subject into the model M1 corresponding to the prediction target “position (promising promotion)”. Specifically, the prediction unit 133 inputs the behavior information ADT1 into the model M1. The model M1 to which the action information ADT1 is input outputs a score. In the example of FIG. 2, the model M1 to which the action information ADT1 is input outputs a score “0.7” as shown in the score SC11-1.

In the example of FIG. 2, the prediction unit 133 predicts the promising degree of the user U1 based on the score output by the model M1. For example, the prediction unit 133 generates prediction information PL11 based on the prediction of the promising degree of the user U1. The prediction unit 133 predicts the promising degree of the user U1 based on the score output by the model M1 and the information stored in the promising degree information storage unit 124. The prediction unit 133 predicts the promising degree of the user U1 based on the information (correspondence information) in which the score and the promising degree are associated with each other as shown in the promising degree information storage unit 124 in FIG.

The prediction unit 133 compares the score output by the model M1 with the score range of the associated information in the promising information storage unit 124 shown in FIG. 2, and the score output by the model M1 falls within any score range. Determine if it is included. In the example of FIG. 2, the prediction unit 133 determines that the score output by the model M1 is “0.7” and corresponds to the score “0.6 or more” corresponding to the promising degree “high”. Therefore, as shown in the prediction information PL11, the prediction unit 133 predicts that the promising degree corresponding to the prediction target “position (promising degree of promotion)” of the user U1 is “high”.

In the example of FIG. 2, the prediction unit 133 predicts the job title of the user U1 after Z to Y years using the rank list RL11. The prediction unit 133 compares the score output by the model M1 with the score of the promotion information in the rank list RL11 shown in FIG. 2, and determines which score or higher the score output by the model M1 is. In the example of FIG. 2, in the prediction unit 133, the score output by the model M1 is "0.7", the score of the B position (manager) is "0.6" or more, and the score of the A position (director) is "0.7". It is determined that it is less than 1 ”. Therefore, as shown in the prediction information PL11, the prediction unit 133 predicts the position of the user U1 in the prediction target “position (promising promotion)” as “manager or higher”. For example, the prediction unit 133 predicts that the position of the user U1 after Z to Y years is "manager or higher". In this case, the prediction unit 133 predicts that the user U1 will be promoted to a position of manager or higher in the years Z to Y. For example, the forecasting department 133 predicts that in 2022, five years after 2017, user U1 will be promoted to a position of manager or higher.

For example, the prediction unit 133 calculates the score using a model having an arbitrary structure such as the regression model and the neural network described above. Specifically, the model M1 is a value obtained by quantifying the prediction of a predetermined target (that is, the action subject) when the user's behavior information (that is, each element used for calculating the score described above) is input. The coefficient is set to output a score that suggests whether the prediction target is highly promising. The prediction unit 133 uses such a model M1 to calculate a score regarding the promising degree of the action subject.

In the above example, the model M1 is a model that outputs a quantified value of the prediction of the promising degree of the action subject when the action information of the user is input. However, the model (model X) according to the embodiment may be a model generated based on the result obtained by repeating the input / output of data to the model M1. For example, the model X may be a model (model Y) trained to input the user's behavior information and output the score output by the model M1. Alternatively, the model M1 may be a model trained to input the user's behavior information and output the output value of the model Y. Further, when the prediction unit 133 performs the prediction processing using GAN (Generative Adversarial Networks), the model M1 may be a model forming a part of GAN.

(Providing section 134)
The providing unit 134 provides various information. For example, the providing unit 134 provides various information to the terminal device 10. For example, the providing unit 134 provides a service regarding the promising degree of the action subject predicted by the predicting unit 133. For example, the providing unit 134 provides the user with information on the promising degree of the action subject predicted by the prediction unit 133. For example, the providing unit 134 provides the terminal device 10 used by the user with information on the promising degree of the action subject predicted by the predicting unit 133. Further, for example, the providing unit 134 may provide information about the model generated by the generating unit 132 to an external information processing device. Further, for example, the providing unit 134 may provide the information output by the model to an external information processing device.

For example, the providing unit 134 provides information based on the predicted promising degree of the action subject. In the example of FIG. 2, the providing unit 134 provides the prediction information PL11 of the user U1 to a company that desires information regarding the promising degree of the user U1. The providing unit 134 provides the prediction information PL11 of the user U1 to the business terminal 20 used by the administrator MA1 of the company CP11 in which the user U1 is undergoing the recruitment test. The providing unit 134 transmits the prediction information PL11 of the user U1 to the business terminal 20 used by the administrator MA1.

[4. Flow of generation process]
Next, the procedure of the generation process by the prediction system 1 according to the embodiment will be described with reference to FIG. FIG. 10 is a flowchart showing an example of the generation process according to the embodiment.

As shown in FIG. 10, the prediction device 100 acquires learning data (step S101). For example, the prediction device 100 acquires learning data from the learning data storage unit 121.

After that, the prediction device 100 generates a model based on the training data (step S102). In the example of FIG. 1, the prediction device 100 generates a model M1 or the like from the learning data storage unit 121 using the learning data.

[5. Prediction processing flow]
Next, the procedure of the generation process by the prediction system 1 according to the embodiment will be described with reference to FIG. FIG. 11 is a flowchart showing an example of the prediction process according to the embodiment.

As shown in FIG. 11, the prediction device 100 acquires the user's behavior information (step S201). In the example of FIG. 2, the prediction device 100 acquires the behavior information ADT1 of the user U1.

Further, the prediction device 100 predicts the promising degree of the action subject using the user's action information and the model (step S202). In the example of FIG. 2, the prediction device 100 predicts the promising degree of the position of the user U1 based on the score output by the model M1.

In addition, the prediction device 100 provides information on the predicted promising degree of the user (step S203). In the example of FIG. 2, the prediction device 100 provides the prediction information PL11 to the business terminal 20 used by the manager MA1 of the company CP11 in which the user U1 is undergoing the recruitment test.

[6. effect〕
As described above, the prediction device 100 according to the embodiment has an acquisition unit 131 and a prediction unit 133. The acquisition unit 131 acquires action information indicating a history of actions by a predetermined action subject. Further, the prediction unit 133 predicts the promising degree of the future potential of a predetermined action subject at a time point after the action based on the action information acquired by the acquisition unit 131.

As described above, the prediction device 100 according to the embodiment predicts the promising degree of the future potential of the predetermined action subject at a time point after the action based on the action information indicating the history of the action by the predetermined action subject. Therefore, the promising degree of the action subject can be appropriately predicted.

Further, in the prediction device 100 according to the embodiment, the acquisition unit 131 acquires a model that outputs a score indicating the promising degree of the predetermined action subject in response to the input of the action information of the predetermined action subject. The prediction unit 133 inputs the behavior information into the model, and predicts the promising degree of a predetermined behavior subject based on the score output by the model.

In this way, the prediction device 100 according to the embodiment inputs the behavior information into the model that outputs the score indicating the promising degree of the predetermined behavior subject in response to the input of the behavior information of the predetermined behavior subject, and the model outputs the behavior information. By predicting the promising degree of a predetermined action subject based on the score obtained, the promising degree of the action subject can be appropriately predicted.

Further, in the prediction device 100 according to the embodiment, the acquisition unit 131 acquires the action information of a predetermined action subject who is a user.

In this way, the prediction device 100 according to the embodiment appropriately obtains the behavior information of a predetermined action subject who is the user and predicts the promising degree of the user, thereby appropriately determining the promising degree of the user who is a human (natural person). Can be predicted.

Further, in the prediction device 100 according to the embodiment, the acquisition unit 131 acquires behavior information including a history of user behavior on the Internet.

As described above, the prediction device 100 according to the embodiment appropriately determines the promising degree of the user who is a human (natural person) by predicting the promising degree of the user by using the behavior information including the history of the user's behavior on the Internet. Can be predicted.

Further, in the prediction device 100 according to the embodiment, the acquisition unit 131 acquires behavior information including a search history by the user.

As described above, the prediction device 100 according to the embodiment appropriately predicts the promising degree of the user who is a human (natural person) by predicting the promising degree of the user by using the behavior information including the search history by the user. be able to.

Further, in the prediction device 100 according to the embodiment, the acquisition unit 131 acquires behavior information including a purchase history by the user.

As described above, the prediction device 100 according to the embodiment appropriately predicts the promising degree of the user who is a human (natural person) by predicting the promising degree of the user by using the behavior information including the purchase history by the user. be able to.

Further, in the prediction device 100 according to the embodiment, the prediction unit 133 predicts the promising degree of the user's job title.

As described above, the prediction device 100 according to the embodiment can predict whether the user will be promoted in an organization such as a company by predicting the promising degree of the user's job title, and thus appropriately determine the promising degree of the user. Can be predicted.

Further, in the prediction device 100 according to the embodiment, the acquisition unit 131 acquires the action information of the action by a predetermined action subject which is a predetermined group.

In this way, the prediction device 100 according to the embodiment acquires behavioral information of actions by a predetermined action subject, which is a predetermined group, and predicts the promising degree of the predetermined group, for example, by a plurality of user groups or the like. It is possible to appropriately predict the promising degree of a predetermined organization such as a constituent organization.

Further, in the prediction device 100 according to the embodiment, the acquisition unit 131 acquires the action information of the action by the predetermined action subject which is a predetermined corporation.

In this way, the prediction device 100 according to the embodiment acquires behavioral information of actions by a predetermined action subject, which is a predetermined corporation, and predicts the promising degree of the predetermined corporation, for example, by an action subject having legal personality. It is possible to appropriately predict the promising degree of a certain corporation.

[7. Hardware configuration]
The prediction device 100 according to the above-described embodiment is realized by, for example, a computer 1000 having a configuration as shown in FIG. FIG. 12 is a hardware configuration diagram showing an example of a computer that realizes the function of the prediction device. The computer 1000 includes a CPU 1100, a RAM 1200, a ROM (Read Only Memory) 1300, an HDD (Hard Disk Drive) 1400, a communication interface (I / F) 1500, an input / output interface (I / F) 1600, and a media interface (I / F). ) Has 1700.

The CPU 1100 operates based on a program stored in the ROM 1300 or the HDD 1400, and controls each part. The ROM 1300 stores a boot program executed by the CPU 1100 when the computer 1000 is started, a program that depends on the hardware of the computer 1000, and the like.

The HDD 1400 stores a program executed by the CPU 1100, data used by such a program, and the like. The communication interface 1500 receives data from another device via the network N and sends it to the CPU 1100, and transmits the data generated by the CPU 1100 to the other device via the network N.

The CPU 1100 controls an output device such as a display or a printer, and an input device such as a keyboard or a mouse via the input / output interface 1600. The CPU 1100 acquires data from the input device via the input / output interface 1600. Further, the CPU 1100 outputs the generated data to the output device via the input / output interface 1600.

The media interface 1700 reads a program or data stored in the recording medium 1800 and provides the program or data to the CPU 1100 via the RAM 1200. The CPU 1100 loads the program from the recording medium 1800 onto the RAM 1200 via the media interface 1700, and executes the loaded program. The recording medium 1800 is an optical recording medium such as a DVD (Digital Versatile Disc) or PD (Phase change rewritable Disk), a magneto-optical recording medium such as an MO (Magneto-Optical disk), a tape medium, a magnetic recording medium, or a semiconductor memory. And so on.

For example, when the computer 1000 functions as the prediction device 100 according to the embodiment, the CPU 1100 of the computer 1000 is controlled by executing a program or data (for example, model M1 (model data MDT1)) loaded on the RAM 1200. The function of the unit 130 is realized. The CPU 1100 of the computer 1000 reads and executes these programs or data (for example, model M1 (model data MDT1)) from the recording medium 1800, but as another example, these programs from another device via the network N. May be obtained.

Although some of the embodiments and modifications of the present application have been described in detail with reference to the drawings, these are examples and vary based on the knowledge of those skilled in the art, including the embodiments described in the disclosure line of the invention. It is possible to carry out the present invention in another form obtained by modifying or improving the above.

[8. Others]
Further, among the processes described in the above-described embodiments and modifications, all or part of the processes described as being automatically performed can be manually performed, or the processes described as being manually performed. It is also possible to automatically perform all or part of the processed processing by a known method. In addition, the processing procedure, specific name, and information including various data and parameters shown in the above document and drawings can be arbitrarily changed unless otherwise specified. For example, the various information shown in each figure is not limited to the illustrated information.

Further, each component of each of the illustrated devices is a functional concept, and does not necessarily have to be physically configured as shown in the figure. That is, the specific form of distribution / integration of each device is not limited to the one shown in the figure, and all or part of the device is functionally or physically distributed / physically in any unit according to various loads and usage conditions. It can be integrated and configured.

In addition, the above-described embodiments and modifications can be appropriately combined as long as the processing contents do not contradict each other.

In addition, the above-mentioned "section, module, unit" can be read as "means" or "circuit". For example, the acquisition unit can be read as an acquisition means or an acquisition circuit.

1 Prediction system 100 Prediction device 121 Learning data storage unit 122 Model information storage unit 123 User information storage unit 124 Promising information storage unit 125 Prediction information storage unit 130 Control unit 131 Acquisition unit 132 Generation unit 133 Prediction unit 134 Providing unit 10 Terminal device N network

Claims (13)

And action information indicating a history of behavior by a given actors, an acquisition unit that acquires the potential of a list including a plurality of potential level, each associated with a range of scores,
According to the comparison between the score based on the action information acquired by the acquisition unit and the promising degree list, the promising degree regarding the future potential of the predetermined action subject at a time point after the action is promising. A predictor that predicts which of the degrees is promising,
A prediction device characterized by comprising. An acquisition unit that acquires action information showing a history of actions including a search by a predetermined action subject who is a user, and a list of promising degrees including a plurality of promising degrees each associated with a range of scores .
And scores rather based on behavior information acquired by the acquisition unit, in accordance with the comparison between the potential of the list, promising degree on the future of the predetermined actors at a later point in time than the action, said plurality A prediction unit that predicts which of the promising degrees of
A prediction device characterized by comprising. The acquisition unit
The prediction device according to claim 1, wherein the behavior information of the predetermined behavioral subject who is a user is acquired. The acquisition unit
The prediction device according to claim 2, wherein the behavior information including the history of the behavior of the user on the Internet is acquired. The acquisition unit
The prediction device according to claim 3, wherein the behavior information including the search history by the user is acquired. The acquisition unit
The prediction device according to any one of claims 2 to 5, wherein the behavior information including the browsing history of the content by the user is acquired. The acquisition unit
The prediction device according to any one of claims 3 to 6, wherein the behavior information including the purchase history by the user is acquired. The prediction unit
The prediction device according to any one of claims 3 to 7, wherein the promising degree regarding the position of the user is predicted. The acquisition unit
In response to the input of the action information of the predetermined action subject, a model for outputting a score indicating the promising degree of the predetermined action subject is acquired.
The prediction unit
The invention according to any one of claims 1 to 8 , wherein the behavior information is input to the model, and the promising degree of the predetermined behavior subject is predicted based on the score output by the model. Predictor. It ’s a computer-based prediction method.
And action information indicating a history of behavior by a given actors, an acquisition step of acquiring a potential degree list containing a plurality of potential level, each associated with a range of scores,
According to the comparison between the score based on the behavior information acquired by the acquisition step and the promising degree list, the promising degree regarding the future potential of the predetermined action subject at a time point after the behavior is the plurality of promising. a prediction step of predicting which one of promising degree of degree,
A prediction method characterized by including. And action information indicating a history of behavior by a given actors, and acquisition procedure, each acquires the potential of a list including a plurality of potential level associated with the range of scores,
According to the comparison between the score based on the action information acquired by the acquisition procedure and the promising degree list, the promising degree regarding the future potential of the predetermined action subject at a time point after the action is the plurality of promising. a prediction procedure for predicting whether it is a promising degree of degree,
A prediction program characterized by having a computer execute. It ’s a computer-based prediction method.
An acquisition process for acquiring behavioral information showing a history of behaviors including a search by a predetermined behavioral subject who is a user, and a promising degree list including a plurality of promising degrees each associated with a score range .
And scores rather based on behavior information acquired by the acquisition step, in response to a comparison between the potential of the list, promising degree on the future of the predetermined actors at a later point in time than the action, said plurality The prediction process for predicting which of the promising degrees of
A prediction method characterized by including. An acquisition procedure for acquiring action information showing a history of actions including a search by a predetermined action subject who is a user, and a list of promising degrees including a plurality of promising degrees each associated with a range of scores .
And scores rather based on behavior information acquired by the acquisition procedure, in accordance with the comparison between the potential of the list, promising degree on the future of the predetermined actors at a later point in time than the action, said plurality Prediction procedure for predicting which of the promising degrees of
A prediction program characterized by having a computer execute.

Images