JP2019139413A - Search processing program, search processing method, and search processing apparatus - Google Patents

Abstract

PROBLEM TO BE SOLVED: To improve search accuracy. SOLUTION: An input receiving unit accepts a search word input by a plurality of users on a search screen, and also receives an evaluation on a search result input by a plurality of users on the search evaluation screen, and the evaluation unit receives a high evaluation. Identify the user as an "intelligence". Then, when the input reception unit accepts the search word input by a user different from the expert on the search screen (for beginners) (S102: affirmative), the search word conversion unit receives the search word based on the search history of the expert. (S106 to S118), and the search execution unit executes a search using the converted search word (S122). [Selection diagram] FIG. 8

Description

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

  Conventionally, when searching for a required document from documents such as reports stored in an in-house database, a user searches the database using a search word that is likely to be included in the document to be searched. Is common. In this case, if the user has abundant work experience, an appropriate search word can be set based on the experience.

JP 2005-301856 A

  However, a user who has little work experience may not be able to set an appropriate search word, and thus may not be able to find a document he needs.

  In one aspect, an object of the present invention is to provide a search processing program, a search processing method, and a search processing device that can improve search accuracy.

  In one aspect, the search processing program receives search conditions corresponding to each of a plurality of users, receives an input of an evaluation regarding a search result corresponding to each of the plurality of users, identifies a user with a high evaluation, When a search condition corresponding to a user different from a user with a high evaluation is received, the search condition is converted based on a search history of the user with a high evaluation, and a search based on the search condition obtained by conversion is executed. A search processing program for causing a computer to execute processing.

  Search accuracy can be improved.

It is a figure showing roughly the composition of the search processing system concerning one embodiment. FIG. 2A is a diagram illustrating the hardware configuration of the server, and FIG. 2B is a diagram illustrating the hardware configuration of the user terminal. It is a functional block diagram of a server. FIG. 4A is a diagram showing a user table, and FIG. 4B is a diagram showing a search history table. Fig.5 (a) is a figure which shows a frequency table, FIG.5 (b) is a figure which shows a TF-IDF table. It is a flowchart which shows the process of the server when a user performs a search using a search screen. Fig.7 (a)-FIG.7 (c) are figures which show the example of a screen displayed on a user terminal, when the process of FIG. 6 is performed. It is a flowchart which shows the process of the server when a user searches using a search screen (for beginners). FIG. 9A to FIG. 9C are diagrams showing examples of screens displayed on the user terminal when the processing of FIG. 8 is performed. Fig.10 (a)-FIG.10 (c) are figures which show another example of a screen example.

  Hereinafter, an embodiment of a search processing system will be described in detail with reference to FIGS. The search processing system according to the present embodiment allows a user to search for documents (electronic data) managed by a company, even if the user is not a user who has a lot of work experience (referred to as an expert). This is a system that supports searching so that the user can search results and search document data.

  FIG. 1 schematically shows a configuration of a search processing system 100 according to the present embodiment. As shown in FIG. 1, the search processing system 100 includes a server 10 as a search processing device and a user terminal 70. The server 10 and the user terminal 70 are connected to a network 80 such as the Internet or a LAN (Local Area Network).

(Server 10)
FIG. 2A shows the hardware configuration of the server 10. As shown in FIG. 2A, the server 10 includes a CPU (Central Processing Unit) 90, a ROM (Read Only Memory) 92, a RAM (Random Access Memory) 94, and a storage unit (here, an HDD (Hard Disk Drive)). 96, a network interface 97, a portable storage medium drive 99, and the like. Each component of the server 10 is connected to a bus 98. In the server 10, the CPU 90 executes a program (including a search processing program) stored in the ROM 92 or the HDD 96, or a program read by the portable storage medium drive 99 from the portable storage medium 91, whereby FIG. The function of each part shown is realized. FIG. 3 also shows a database (database) and a table stored in the HDD 96 of the server 10.

  FIG. 3 shows a functional block diagram of the server 10. As shown in FIG. 3, in the server 10, when the CPU 90 executes the program, the input receiving unit 20, the search word conversion unit 22, the search execution unit 24, the providing unit 26, the evaluation unit 28 as the specifying unit, the update The function as the unit 30 is realized. 3 may be realized by an integrated circuit such as an application specific integrated circuit (ASIC) or a field programmable gate array (FPGA).

  The input receiving unit 20 acquires a search word input by the user on the search screen (see FIG. 7A) or the search screen (for beginners) (see FIG. 9A), and executes the search execution unit 24 or search word conversion. Delivered to the unit 22. Further, the input receiving unit 20 acquires information that a button on the screen is pressed by the user and notifies each unit. Note that the user selects and uses either the search screen of FIG. 7A or the search screen (for beginners) of FIG. 9A when searching for a document.

  The search word conversion unit 22 converts the search word input by the user as necessary, and passes the converted search word to the search execution unit 24. Note that the search word conversion unit 22 executes processing using the synonym dictionary DB 45, the user table 42, the frequency table 46, and the TF-IDF table 48. The synonym dictionary DB 45 is a database that stores synonyms in association with each other, and stores information such as “performance test”, “performance test”, and “usefulness test” as synonyms. Details of the user table 42, the frequency table 46, and the TF-IDF table 48 and details of the processing of the search word conversion unit 22 will be described later.

  The search execution unit 24 searches for a document from an enormous amount of documents stored in the document DB 40 using the search word input by the user or the search word converted by the search word conversion unit 22. In addition, the search execution unit 24 transmits the information of the searched document to the user terminal 70 as a search result. Furthermore, the search execution unit 24 stores the search word used for the search in the search history table 44 when an expert performs a search using the search screen of FIG.

  Here, the search execution unit 24 determines, based on the user table 42, whether the user who has searched using the search screen of FIG. FIG. 4A shows the table structure of the user table 42. The user table 42 stores information on “user”, “evaluation average value”, “classification”, and “department” in association with each other. The “user” column stores the name and identification number of the user. The “evaluation average value” column stores an average value of the points evaluated by the user for his / her own search. The “classification” column stores the classification of each user determined from the average evaluation value of each user. There are two categories: “intelligence” and “general”. “Intelligence” means, for example, a user whose evaluation average value is greater than or equal to the threshold and has abundant work experience, and “general” means a user whose evaluation average value is less than the threshold, for example, and whose work experience is not so rich It shall be. In addition, it is good also as setting not only this but the user of a high-order predetermined ratio from the one where evaluation average value is high among all users as an expert. In the “department” column, information on departments in the company to which each user belongs, and information on the accounting department, business department, department A, department B, and the like is stored.

  FIG. 4B shows the table structure of the search history table 44. As shown in FIG. 4B, the search history table 44 stores “intelligence” and information of “word 1”, “word 2”,. In the “intelligence” column, the names and identification numbers of the users who are experts are stored, and in the “word 1”, “word 2”,... Columns, the search words used by the expert in the search are stored. . Note that the words used in one search arranged in the column direction in the search history table 44 are also referred to as “search word group”.

  When the user selects one document from the search results of the search execution unit 24, the providing unit 26 transmits information indicating an outline of the selected document to the user terminal 70. In addition, when a request for downloading a document is received from the user, the providing unit 26 reads out the document requested to be downloaded from the document DB 40 and transmits (downloads) the document to the user terminal 70.

  When the user downloads a document on the user terminal 70, the evaluation unit 28 inquires the user about the evaluation regarding the current search, and updates the user table 42 based on the evaluation result. In FIG. 3, a broken line means “update” of the table.

  The update unit 30 updates the frequency table 46 based on the data of the user table 42 and the search history table 44, and the TF-IDF table 48 based on the data stored in the search history table 44 and the document DB 40. Update. Note that the update unit 30 executes update processing of each table by batch processing at a frequency of once a day or once a week.

FIG. 5A shows the table structure of the frequency table 46. As shown in FIG. 5A, the frequency table 46 includes information (appearance frequency) on how often each search word (“solvent”, “acetone”, etc.) is used for the search by experts in each department. Stored. For example, the appearance frequency F _Aα of the word α in the A section is expressed as follows, where M _{Aα is} the number of searches performed by the expert in the A section using the word α and N _A is the total number of searches performed by the expert in the A section. It can be represented by Formula (1).
_{F Aα = M Aα / N A} ... (1)

  The update unit 30 identifies experts in each department based on the user table 42, refers to the search history table 44, calculates the appearance frequency of each word for each department, and updates the frequency table 46.

  FIG. 5B shows the table structure of the TF-IDF table 48. As shown in FIG. 5B, the TF-IDF table 48 includes “TF (Term Frequency) value”, “IDF (Inverse Document Frequency)” for each search word (“solvent”, “acetone”, etc.). Each value of “value”, “TF-IDF value”, and “distribution of TF-IDF value” is stored. Here, the “TF value” means the frequency at which each search word appears in all documents stored in the document DB 40. The “IDF value” is a value indicating how many documents each search word is used in common, and the smaller the value, the less the document appears in the document. The “TF-IDF value” is a product of the above-described TF value and the above-described IDF value. The dispersion of the TF-IDF value is a value indicating how much the TF-IDF value for each document, which is the product of the TF value obtained for each document for a certain search word and the IDF value, varies. A search word having a small variance in TF-IDF values indicates that the word appears in the same degree in any document, and a search word having a greater variance in TF-IDF values contributes to document classification. It means that.

  The update unit 30 specifies the search word used for the search based on the search history table 44, calculates each value of the TF-IDF table 48 using the document DB 40 for each specified search word, and updates it. To do.

(User terminal 70)
Returning to FIG. 1, the user terminal 70 is a terminal used by users in the company. The user can browse and use the document searched by the server 10 by inputting a search word to the user terminal 70 and performing a predetermined operation.

  FIG. 2B shows the hardware configuration of the user terminal 70. As shown in FIG. 2B, the user terminal 70 is stored in the CPU 190, ROM 192, RAM 194, storage unit (HDD) 196, network interface 197, display unit 193, input unit 195, and portable storage medium 191. A portable storage medium drive 199 capable of reading information is provided. The display unit 193 includes a liquid crystal display or the like, and the input unit 195 includes a mouse, a keyboard, a touch panel, or the like. Each component of the user terminal 70 is connected to the bus 198.

(About processing of server 10)
Next, the processing of the server 10 will be described in detail along the flowcharts of FIGS. 6 and 8 with reference to other drawings as appropriate.

  FIG. 6 shows the processing of the server 10 when the user performs a search using the search screen of FIG. 7A, and FIG. 8 shows the search screen (for beginners) of FIG. 9A. The processing of the server 10 in the case of performing a search using) is shown.

(Processing when searching using the search screen)
First, based on FIG. 6, the process of the server 10 when a user performs a search using a search screen is demonstrated. As a premise of the processing in FIG. 6, it is assumed that the input reception unit 20 transmits a search screen as illustrated in FIG. 7A to the user terminal 70 in response to a request from the user. Therefore, it is assumed that the search screen of FIG. 7A is displayed on the display unit 193 of the user terminal 70. The search screen of FIG. 7A includes a field for inputting a search word, a “search” button, and a field for displaying a search result.

  In the process of FIG. 6, first, in step S <b> 10, the input receiving unit 20 waits until a search word is input on the search screen and the “Search” button is pressed. As shown in FIG. 7A, for example, when the user inputs the search word group “solvent plant utility” and presses the “search” button, the process proceeds to step S12. When the “search” button is pressed, the search word (search word group) input from the user terminal 70 is transmitted, so that the input receiving unit 20 sends the received search word to the search execution unit 24. Deliver.

  In step S 12, the search execution unit 24 searches for a document in the document DB 40 using the input search word, and transmits the search result to the user terminal 70. Thereby, the search result is displayed on the display unit 193 of the user terminal 70. FIG. 7A shows a state where information of three documents is displayed as a search result.

  Next, in step S <b> 14, the input reception unit 20 waits until one of the search results is selected on the user terminal 70. When the user selects one document from the search results on the user terminal 70, the input receiving unit 20 notifies the providing unit 26 to that effect and proceeds to step S16.

  In step S <b> 16, the providing unit 26 transmits a screen showing an outline of the selected document to the user terminal 70. As a result, a screen for displaying an outline of the document as shown in FIG. 7B is displayed on the display unit 193 of the user terminal 70. It is assumed that a “download” button is also displayed on the screen of FIG.

  Next, in step S <b> 18, the input reception unit 20 stands by until the “download” button is pressed by the user at the user terminal 70. When the “download” button is pressed by the user at the user terminal 70, the input receiving unit 20 notifies the providing unit 26 to that effect, and the process proceeds to step S20.

  If transfering it to step S20, the provision part 26 will read the data of the document selected by the user from document DB40, and will transmit to the user terminal 70 (it downloads).

  After step S20, the process proceeds to step S22, and the evaluation unit 28 transmits a search evaluation screen to the user terminal 70. As a result, a search evaluation screen as shown in FIG. 7C is displayed on the display unit 193 of the user terminal 70. In addition, a search evaluation screen is a screen which evaluates the satisfaction with respect to the search result which the user implemented, for example in 11 steps (0-10 points). The user can transmit the evaluation result to the server 10 (input reception unit 20) by selecting the score and pressing the “Send” button.

  Next, in step S <b> 24, the input receiving unit 20 stands by until the “Send” button is pressed on the user terminal 70. When the “send” button is pressed by the user, the input reception unit 20 notifies the evaluation unit 28 of the number of points transmitted from the user terminal 70, and the process proceeds to step S26.

  In step S26, the evaluation unit 28 calculates an average value of the evaluation points selected (input) by the user and the evaluation points obtained so far, and the “evaluation average value” of the user table 42 is calculated. "" Column is updated. Note that the evaluation unit 28 also changes the category when it becomes necessary to change the category of the user due to the update of the evaluation average value. For example, if the user needs to satisfy the criterion that “the evaluation average value is equal to or greater than a predetermined threshold” in order to become an “informed person”, the evaluation unit 28 determines that the updated evaluation average value is equal to or greater than the threshold value. Whether or not and the category is updated based on the determination result.

  Next, in step S <b> 28, the search execution unit 24 updates the search history table 44 with the input search word when the user who performed the search is an “informed person”. More specifically, the search execution unit 24 determines whether or not the user is an expert with reference to the user table 42. If the user is an expert, the search input on the screen of FIG. 7A. The word group “solvent plant usefulness” and user identification information are added to the search history table 44.

  Thus, all the processes in FIG. 6 are completed. Note that the processing in FIG. 6 is repeatedly executed every time a search request using the search screen in FIG.

(Process when searching using the search screen (for beginners))
Next, based on FIG. 8, the process of the server 10 when a user performs a search using a search screen (for beginners) will be described. As a premise of the processing in FIG. 8, the input receiving unit 20 transmits a search screen (for beginners) as shown in FIG. 9A to the user terminal 70 in response to a request from the user. To do. Accordingly, it is assumed that the search screen (for beginners) in FIG. 9A is displayed on the display unit 193 of the user terminal 70. The screen shown in FIG. 9A includes a “synonym conversion” button and a “convert from history” button in addition to a “search” button. It is assumed that the “convert from history” button cannot be pressed (cannot be selected) until the “synonym conversion” button is pressed.

  In step S102 of FIG. 8, the input receiving unit 20 waits until a search word is input on the search screen (for beginners). As shown in FIG. 9A, for example, when the user inputs the search word group “acetone overseas performance test”, the process proceeds to step S104.

  In step S104, the input receiving unit 20 determines whether the “synonym conversion” button has already been pressed. If the determination in step S104 is negative, the process proceeds to step S106.

  In step S106, the input receiving unit 20 determines whether the “synonym conversion” button has been pressed. If the determination in step S106 is negative, the process proceeds to step S120, and the input receiving unit 20 determines whether or not the “search” button has been pressed. If the determination in step S120 is negative, the process returns to step S104. That is, the determinations in steps S104, S106, and S120 are repeatedly executed until the determination in step S106 or step S120 is affirmed. If the user wants to perform a search using the search word group “acetone overseas performance test” shown in FIG. 9A, he / she presses the “search” button and converts the search word into an appropriate word. Press the “Convert from history” button.

  When the user presses the “search” button, the determination in step S120 is affirmed, and the process proceeds to step S122. In step S122, the search execution unit 24 searches the document DB 40 using the input search word group “acetone overseas performance test”, and displays the search result on the search screen (for beginners).

  On the other hand, when the user presses the “synonym conversion” button, the determination in step S106 is affirmed, and the process proceeds to step S108. If transfering it to step S108, the search word conversion part 22 will extract the synonym of each search word from the synonym dictionary DB45. For example, when the input search word is the search word group “acetone overseas performance test”, the synonym dictionary DB 45 includes information that “performance test”, “performance test”, and “usefulness test” are synonyms. Suppose it was. In this case, the search word conversion unit 22 extracts “performance test” and “usefulness test” as synonyms of “performance test”. In addition, when synonyms exist in “acetone” or “overseas”, they are extracted in the same manner, but it is assumed that no synonyms exist for the sake of simplicity of explanation.

  Next, in step S110, the search word conversion unit 22 specifies the appearance frequency of each synonym (including the search word) in the search for the expert in the same department. Specifically, when the search word conversion unit 22 refers to the user table 42 and specifies that the user performing the search is, for example, the A department, the search word conversion unit 22 refers to the frequency table 46 to determine the user of the A department. In the search history, the appearance frequency of “performance test”, “performance test”, and “usefulness test” is specified.

  Next, in step S112, the search word conversion unit 22 sets a search word having the highest appearance frequency among the synonyms (including the search word). In the above example, if the appearance frequency of “performance test” is 0.1, the appearance frequency of “performance test” is 0.2, and the appearance frequency of “usefulness test” is 0.3, the appearance frequency is the highest. “Usefulness test” is the search word. That is, as shown in FIG. 9B, the search word conversion unit 22 converts the search word group from the “acetone overseas performance test” in FIG. 9A to the “acetone overseas usefulness test”. As a result, the input search word can be converted into a search word suitable for a search frequently used in the department without changing the content that the user wants to search. When synonyms exist in a plurality of search words, conversion based on the appearance frequency is performed for each.

  Next, in step S114, the input receiving unit 20 determines whether or not the “convert from history” button has been pressed. If the determination in step S114 is negative, the process proceeds to step S120, and the input receiving unit 20 determines whether or not the “search” button has been pressed. If the determination in step S120 is negative, the process returns to step S104. However, since the determination in step S104 is affirmative, the process proceeds to step S114 again. Therefore, the input receiving unit 20 waits until the determination of either step S114 or S120 is affirmed. The user presses the “Search” button when searching for the search word group “acetone overseas usability test” shown in FIG. 9B, and selects “history” when the search word is further converted. Press the “Convert from” button.

  When the user presses the “search” button, the determination in step S120 is affirmed, and the process proceeds to step S122. In step S122, the search execution unit 24 searches the document DB 40 using the search word group “acetone overseas usability test” shown in FIG. 9B, and displays the search result.

  On the other hand, when the user presses the “convert from history” button, the determination in step S114 is affirmed, and the process proceeds to step S116. In step S116, the search word conversion unit 22 fixes search words other than one search word. In this case, the search word conversion unit 22 fixes search words other than one search word as a search axis, and sets one search word that is not fixed as a conversion target. For example, comparing “acetone” with “usefulness test”, “acetone” is more characteristic and is a search word that is more important and useful in document search, so it should not be excluded from the search word . Therefore, in this embodiment, the search word conversion unit 22 uses the IDF value of each word stored in the TF-IDF table 48 in order to fix a useful search word.

  For example, when the search word group is “acetone overseas usefulness test”, the search word conversion unit 22 refers to the TF-IDF table 48 and fixes the words other than the smallest IDF value among the three search words. . In this way, words such as proper nouns (words whose IDF values tend to increase) that are not used universally (not widely used in many documents) can be fixed as search words. In the above example, “Acetone” and “Overseas” are fixed in the “Acetone Overseas Usefulness Test”.

  Next, in step S118, the search word conversion unit 22 converts the search words that are not fixed based on the search history of experts in the same department as the user. Specifically, the search word conversion unit 22 refers to the search history table 44, and searches the search history (search word group) including all the search words fixed in step S116 from the search history of the expert in the same department as the user. Extract. Here, for example, “acetone overseas use restriction”, “acetone overseas use example”, and “acetone overseas use example China” are extracted. In this case, the words “Usage restrictions”, “Usage examples”, and “Usage examples China” other than “Acetone overseas” are candidate words (conversion candidate words) to be converted to “usefulness test”.

  Next, the search word conversion unit 22 refers to the TF-IDF table 48 and extracts the variance of the TF-IDF values of words (conversion candidate words) other than the fixed search word. If the conversion candidate word is “use restriction”, the distribution of the “use restriction” TF-IDF value is extracted from the TF-IDF table 48. If the conversion candidate word is “use example”, the distribution of the TF-IDF values of “use example” is extracted from the TF-IDF table 48. Further, if the conversion candidate word is “use example China”, the variance of the TF-IDF value of “use example” and the variance of the TF-IDF value of “China” are extracted from the TF-IDF table 48, and The average value is the variance of the TF-IDF values of the conversion candidate word “Usage example China”.

  Then, the search word conversion unit 22 uses the “use restriction”, “use example”, and “use example China” instead of the search word that does not fix the one having the largest variance of the TF-IDF values. To do. As described above, the reason why a word having a large TF-IDF value variance is used as a search word is that a word having a larger TF-IDF value variance is considered to be a word that contributes to document classification. In the above example, for example, if the TF-IDF value variance of “Usage Restriction” is the largest, the search word conversion unit 22 performs the search word group “acetone overseas” in FIG. 9B as shown in FIG. Convert “usefulness test” to “restricted use of acetone overseas”. Thereafter, the process proceeds to step S120.

  After the process proceeds to step S120 through steps S116 and S118, when the “search” button is pressed by the user, the process proceeds to step S122, and the search execution unit 24 uses the search word group “acetone overseas use shown in FIG. 9C”. The document DB 40 is searched using “restriction”, and the search result is displayed.

  After step S122, the process proceeds to step S124, and the input receiving unit 20 waits until one of the search results is selected on the user terminal 70, as in step S14 of FIG. When the user selects one document from the search results on the user terminal 70, the input receiving unit 20 notifies the providing unit 26 to that effect, and the process proceeds to step S126. In step S126, the providing unit 26 transmits a screen showing an outline of the selected document to the user terminal 70, as in step S16 of FIG. 6 (see FIG. 7B).

  Next, in step S128, as in step S18 of FIG. 6, the input receiving unit 20 stands by until the user presses the “download” button on the screen of FIG. When the “download” button is pressed by the user at the user terminal 70, the input receiving unit 20 notifies the providing unit 26 to that effect, and the process proceeds to step S130. When the process proceeds to step S130, the providing unit 26 executes download by reading the data of the selected document from the document DB 40 and transmitting it to the user terminal 70, as in step S20 of FIG.

  Thus, all the processes in FIG. 8 are completed. The process of FIG. 8 is repeatedly executed every time a search request using the search screen (for beginners) of FIG. 9A is issued from the user.

  In the present embodiment, as described above, when the “convert from history” button is pressed, the search word conversion unit 22 displays the result of converting an unfixed search word on the search screen (for beginners). The case of displaying was explained. However, the present invention is not limited to this. For example, it is assumed that the “Convert from history” button is pressed in a state where “acetone overseas usefulness test” is set as a search word group as shown in FIG. At this time, when a plurality of conversion candidate words for the search word “usefulness test” that are not fixed are extracted, as shown in FIG. 10B, the search word group after conversion with each conversion candidate word May be displayed in a selectable manner. The user can change the search word group used for the search as shown in FIG. 10C by selecting one of the displayed search word groups (for example, “acetone overseas use restriction”). In addition, in FIG.10 (b), for example, it can display from the top in an order from a thing with large dispersion | distribution of the TF-IDF value of the word of a conversion candidate. As a result, the search word group that the search word conversion unit 22 determines to be appropriate can be displayed higher (easy to be noticed by the user).

  In the present embodiment, it is assumed that the “convert from history” button can be pressed even if there is only one search word. For example, it is assumed that the user presses the “convert from history” button in a state where “acetone” is set as a search word. In this case, the search word conversion unit 22 fixes the search word “acetone” and extracts the search history including the fixed search word “acetone” from the search history of the expert in the same department as the user. Then, the search word conversion unit 22 refers to the TF-IDF table 48, extracts the TF-IDF value variance of words other than the fixed search word “acetone”, and has the largest variance of the TF-IDF value. To the search word. By doing in this way, when there is one search word input by the user, it becomes possible to perform a search after adding an appropriate search word.

  As is clear from the above description, in this embodiment, the input reception unit 20 causes the search condition reception unit to receive search conditions corresponding to each of a plurality of users, and the search result corresponding to each of the plurality of users. A function as an evaluation receiving unit that receives an input of an evaluation regarding is realized. In addition, when the search word conversion unit 22 and the search execution unit 24 accept search conditions corresponding to a user different from the user with high evaluation, the search condition is converted based on the search history of the user with high evaluation, A function as a search unit that executes a search based on a search condition obtained by conversion is realized.

  As described above in detail, according to the present embodiment, the input receiving unit 20 receives search words input by a plurality of users on the search screen (S10: Yes), and a plurality of users input on the search evaluation screen. The evaluation related to the retrieved result is accepted (S24: affirmative), and the evaluation unit 28 identifies the accepted user who has a high evaluation as an “informed person” (S26). And when the input reception part 20 receives the search word which the different user from the expert input in the search screen (for beginners) (S102: affirmation), the search word conversion part 22 is based on the search history of an expert. The search word is converted (S106 to S118), and the search execution unit 24 executes a search using the converted search word (S122). Thereby, in this embodiment, since the search can be executed after converting the search word input by the user other than the expert into a more appropriate search word based on the search history of the expert, the search accuracy is improved. It becomes possible. In addition, the user does not have to repeat the work of changing a part of the search word and executing the search until the necessary document is found, thereby reducing the labor required for the search and reducing the time required for the search. Can do. In addition, since the search word is automatically converted, the user can easily know what search word to set if it is an expert. As a result, it is possible to improve the user's business ability.

  In the present embodiment, the search word conversion unit 22 uses a search history of an expert in the same department as the user who performs the search in steps S110 and S118. Thereby, the search word set by the user can be converted into an appropriate search word based on the search history of an expert having knowledge to be referred to by the user.

  In this embodiment, the search word conversion unit 22 extracts a synonym (similar word) of each search word from the synonym dictionary DB 45 when the “synonym conversion” button is pressed by the user, and the search word and its synonyms are extracted. Among them, a search word having the highest appearance frequency in the search history of an expert in the same department is set as a search word. As a result, the input search word can be converted into a search word suitable for a search frequently used in the department without changing the content that the user wants to search.

  Further, in this embodiment, when the “convert from history” button is pressed by the user, the search word conversion unit 22 fixes search words other than one search word based on the IDF value and sets it as a search axis. (S116). This makes it possible to convert a search word that is not fixed to an appropriate search word while fixing the direction of search.

  Further, in the present embodiment, the search word conversion unit 22 specifies a search word conversion candidate word that is not fixed from the search history table 44 based on the fixed search word, and specifies the specified conversion candidate word. Among them, the TF-IDF value variance is the largest and the search word that is not fixed is converted (S118). Thereby, the search word can be converted with an appropriate word contributing to the classification of the document among the conversion candidate words.

  In the above embodiment, as shown in FIG. 9B, the search word is fixed when the “convert from history” button is pressed from the state where “acetone overseas usability test” is set as the search word group. As an example of the method, the example using the IDF value has been described. However, the present invention is not limited to this. For example, when the search word is fixed, the search word conversion unit 22 may further refer to the technical term dictionary DB storing technical terms. Specifically, the search word conversion unit 22 fixes a word (for example, acetone) included in the technical term dictionary DB in the search word group “acetone overseas usefulness test”. After that, the search word conversion unit 22 fixes the words other than the search word “overseas” and “usefulness test” other than the smallest IDF value. By doing so, it is possible to improve the search accuracy because the search terms that are not fixed can be converted into more appropriate words while the technical terms are always fixed. Note that the search word may be fixed using only the technical term dictionary DB without using the IDF value.

  In the above embodiment, the case where the appearance frequency and the search history for each department are used in steps S110 and S118 has been described. However, the present invention is not limited to this. That is, the appearance frequency and search history of the entire company may be used without considering the department.

  In the above embodiment, a case has been described in which the search word other than the search word having the smallest IDF value is fixed in step S116. However, the present invention is not limited to this. It may be fixed. Further, an upper limit of the number of search words to be fixed may be determined.

  In the above embodiment, the search word conversion unit 22 uses the dispersion of TF-IDF values when converting the search word in step S116. However, the present invention is not limited to this. For example, the search word conversion unit 22 may use variance of TF values obtained for each document in step S116.

  In the above-described embodiment, the case has been described in which “intelligence” and “general” are automatically distinguished based on the average evaluation value stored in the user table 42, but the present invention is not limited to this. For example, the administrator or the like may distinguish “intelligence” and “general” by manual input.

  In the above-described embodiment, when there are departments with common handling tasks and similar search trends, these departments may be associated in advance. In this case, even different departments may be handled as the same department in steps S110 and S118.

  In addition, although the said embodiment demonstrated the case where a user can select the screen of Fig.7 (a) and the screen of Fig.9 (a), it is not restricted to this. For example, the input receiving unit 20 determines whether the user who performs the search is “informed person” or “general” based on the user table 42. If the user is “informed person”, the screen of FIG. 7A is displayed on the user terminal 70. If the transmission is “general”, the screen in FIG. 9A may be transmitted to the user terminal 70. Further, only the screen shown in FIG. 9A may be prepared as the search screen. In this case, when the user presses the “search” button on the screen of FIG. 9A without pressing the “synonym conversion” button or the “convert from history” button, the processing of FIG. 6 is performed. When the user presses the “synonym conversion” button or the “convert from history” button, the processing after step S106 in FIG. 8 may be performed.

  In the above embodiment, the case where the search processing system 100 is used in a company has been described. However, the present invention is not limited to this. It may be used not only in a company but also in a school or the like, or may be used in a predetermined group. Moreover, it is good also as not limiting the utilization range of the search processing system 100 to a company, a school, etc.

  In the above embodiment, a case has been described in which a search word used when searching for a document is converted based on a search history of a user who has a high evaluation regarding the search result. However, the present invention is not limited to this. That is, when searching for some information, the input search condition may be converted based on a search history of a user who has a high evaluation regarding the search result.

  The above processing functions can be realized by a computer. In that case, a program describing the processing contents of the functions that the processing apparatus should have is provided. By executing the program on a computer, the above processing functions are realized on the computer. The program describing the processing contents can be recorded on a computer-readable recording medium (except for a carrier wave).

  When the program is distributed, for example, it is sold in the form of a portable recording medium such as a DVD (Digital Versatile Disc) or a CD-ROM (Compact Disc Read Only Memory) on which the program is recorded. It is also possible to store the program in a storage device of a server computer and transfer the program from the server computer to another computer via a network.

  The computer that executes the program stores, for example, the program recorded on the portable recording medium or the program transferred from the server computer in its own storage device. Then, the computer reads the program from its own storage device and executes processing according to the program. The computer can also read the program directly from the portable recording medium and execute processing according to the program. Further, each time the program is transferred from the server computer, the computer can sequentially execute processing according to the received program.

  The above-described embodiment is an example of a preferred embodiment of the present invention. However, the present invention is not limited to this, and various modifications can be made without departing from the scope of the present invention.

In addition, the following additional remarks are disclosed regarding description of the above embodiment.
(Appendix 1) Accepting search conditions corresponding to each of a plurality of users,
Receiving an input of an evaluation related to a search result corresponding to each of the plurality of users;
Identify users with high ratings,
When a search condition corresponding to a user different from the user with the high evaluation is received, the search condition is converted based on the search history of the user with the high evaluation, and a search based on the search condition obtained by conversion is executed. To
A search processing program for causing a computer to execute processing.
(Additional remark 2) The search processing program of Additional remark 1 characterized by the user different from the user with said high evaluation being a user engaged in the same business as the user with said high evaluation.
(Supplementary Note 3) The search condition includes a search word,
When converting the search condition, a similar word that is similar to the search word is specified, and the search word that is frequently used for the search among the similar words is prioritized and converted to the search word. ,
The search processing program according to appendix 1 or 2, characterized by:
(Additional remark 4) The said search conditions include the search word group which combined the some search word,
When converting the search condition, a search word not to be converted and a search word to be converted are specified from the search word group, and the specified search word to be converted is based on a search history of a user with a high evaluation. Convert
The search processing program according to any one of appendices 1 to 3, characterized in that:
(Additional remark 5) The process which performs the said search is a process which searches the document containing the said search word group from several documents,
Of the plurality of search words included in the search word group, the one having a small IDF (Inverse Document Frequency) value in the plurality of documents is specified as a search word not subjected to the conversion,
The search processing program according to appendix 4, characterized in that:
(Additional remark 6) When converting the search word of the said conversion object, the search word group containing the search word which does not perform the said conversion is extracted from the search history of the said user with high evaluation, Based on the said search word group extracted, Converting the search word to be converted;
The search processing program according to appendix 5, characterized in that:
(Supplementary Note 7) When converting the search word to be converted, among the search words other than the search word that is not subjected to the conversion included in the extracted search word group, TF (Term Frequency) − in each of the plurality of documents The IDF value having a large variance is preferentially converted with the search target search word.
The search processing program according to appendix 6, characterized in that:
(Appendix 8) Accepting search conditions corresponding to each of a plurality of users,
Receiving an input of an evaluation related to a search result corresponding to each of the plurality of users;
Identify users with high ratings,
When a search condition corresponding to a user different from the user with the high evaluation is received, the search condition is converted based on the search history of the user with the high evaluation, and a search based on the search condition obtained by conversion is executed. To
A search processing method, wherein a computer executes the processing.
(Supplementary Note 9) A search condition receiving unit that receives search conditions corresponding to each of a plurality of users;
An evaluation receiving unit that receives an input of an evaluation related to a search result corresponding to each of the plurality of users;
A specifying unit for specifying a user having a high evaluation;
When a search condition corresponding to a user different from the user with the high evaluation is received, the search condition is converted based on the search history of the user with the high evaluation, and a search based on the search condition obtained by conversion is executed. A search section to
A search processing apparatus comprising:
(Additional remark 10) The search processing apparatus of Additional remark 9 characterized by the user different from the user with said high evaluation being a user engaged in the same business as the user with said high evaluation.
(Supplementary Note 11) The search condition includes a search word,
The search unit identifies a similar word similar to the search word, and converts the search word to the search word with priority given to a high frequency used by the user with a high evaluation among the similar words.
The search processing apparatus according to appendix 9 or 10, characterized in that:
(Supplementary Note 12) The search condition includes a search word group in which a plurality of search words are combined,
The search unit specifies a search word that is not converted from the search word group and a search word to be converted, and converts the specified search word to be converted based on a search history of a user with a high evaluation. ,
The search processing device according to any one of appendices 9 to 11, characterized in that:
(Additional remark 13) The said search part searches the document containing the said search word group from several documents,
Of the plurality of search words included in the search word group, the one having a small IDF value in the plurality of documents is given priority and specified as a search word not subjected to the conversion.
The search processing device according to appendix 12, characterized in that:
(Additional remark 14) The said search part extracts the search word group containing the search word which does not perform the said conversion from the search history of the said user with high evaluation, Based on the extracted said search word group, the search word of the said conversion object Convert,
The search processing device according to supplementary note 13, characterized in that:
(Additional remark 15) The said search part gives priority to the thing with large dispersion | distribution of TF-IDF value in each of these documents among search words other than the search word which does not perform the said conversion contained in the extracted said search word group. To convert the search word to be converted,
15. The search processing device according to appendix 14, characterized in that:

10 server (search processing device)
20 Input reception part (search condition reception part, evaluation reception part)
22 Search word conversion part (part of the search part)
24 Search execution part (part of the search part)
28 Evaluation Department (Specific Part)

Claims (9)

Accept search conditions corresponding to each of multiple users,
Receiving an input of an evaluation related to a search result corresponding to each of the plurality of users;
Identify users with high ratings,
When a search condition corresponding to a user different from the user with the high evaluation is received, the search condition is converted based on the search history of the user with the high evaluation, and a search based on the search condition obtained by conversion is executed. To
A search processing program for causing a computer to execute processing.   The search processing program according to claim 1, wherein the user different from the user with the high evaluation is a user engaged in the same job as the user with the high evaluation. The search condition includes a search word,
When converting the search condition, a similar word that is similar to the search word is specified, and the search word that is frequently used for the search among the similar words is prioritized and converted to the search word. ,
The search processing program according to claim 1 or 2, characterized in that The search condition includes a search word group in which a plurality of search words are combined,
When converting the search condition, a search word not to be converted and a search word to be converted are specified from the search word group, and the specified search word to be converted is based on a search history of a user with a high evaluation. Convert
The search processing program according to any one of claims 1 to 3, wherein: The process of executing the search is a process of searching a document including the search word group from a plurality of documents.
Of the plurality of search words included in the search word group, the one having a small IDF (Inverse Document Frequency) value in the plurality of documents is specified as a search word not subjected to the conversion,
The search processing program according to claim 4, wherein: When converting the search word to be converted, a search word group including a search word that does not perform the conversion is extracted from the search history of the user with high evaluation, and based on the extracted search word group, Convert search words,
The search processing program according to claim 5, wherein: When converting the search word to be converted, the TF (Term Frequency) -IDF value distribution in each of the plurality of documents among the search words included in the extracted search word group other than the search word that is not converted. Priority is given to those having a larger value and conversion with the conversion target search word.
The search processing program according to claim 6. Accept search conditions corresponding to each of multiple users,
Receiving an input of an evaluation related to a search result corresponding to each of the plurality of users;
Identify users with high ratings,
When a search condition corresponding to a user different from the user with the high evaluation is received, the search condition is converted based on the search history of the user with the high evaluation, and a search based on the search condition obtained by conversion is executed. To
A search processing method, wherein a computer executes the processing. A search condition receiving unit that receives search conditions corresponding to each of a plurality of users;
An evaluation receiving unit that receives an input of an evaluation related to a search result corresponding to each of the plurality of users;
A specifying unit for specifying a user having a high evaluation;
When a search condition corresponding to a user different from the user with the high evaluation is received, the search condition is converted based on the search history of the user with the high evaluation, and a search based on the search condition obtained by conversion is executed. A search section to
A search processing apparatus comprising:

Images