CN109697636A - Merchant recommendation method, merchant recommendation device, electronic equipment and medium - Google Patents

Abstract

Embodiments of the present invention relate to the technical field of information processing, and disclose a merchant recommendation method, a merchant recommendation device, an electronic device and a medium. A merchant recommendation method includes: establishing a merchant recommendation model in advance, wherein the merchant recommendation model includes at least a conversion rate model and a return on investment model; determining that the conversion rate model and the return on investment model are The weight in the recommendation model; the rating of the merchant is obtained according to the recommendation model; the recommendation of the merchant is performed according to the rating of the merchant. By adopting the embodiments of the present invention, the merchants recommended for the users are matched with the needs of the users, which helps to improve the user experience and the profit of the merchants.

Description

Merchant recommendation method, merchant recommendation device, electronic device and medium

technical field

The invention relates to the technical field of information processing, and in particular, to a merchant recommendation method, a merchant recommendation device, an electronic device and a medium.

Background technique

With economic development and the wide application of Internet technology, more and more people will choose online e-commerce platforms for shopping in their daily lives. Therefore, how to recommend suitable merchants to users is very important for both users and merchants. important.

However, the inventor found that there are at least the following problems in the related art: since different users have personalized shopping needs, the recommended merchant lists matching different users should also be different. However, many online e-commerce platforms do not have targeted Recommending a merchant list that matches the user's needs for different users results in poor user experience and difficulty in improving the merchant's revenue.

SUMMARY OF THE INVENTION

The purpose of the embodiments of the present invention is to provide a merchant recommendation method, a merchant recommendation device, an electronic device and a medium, so that the merchants recommended for users match the user's needs and help improve user experience and merchant revenue.

In order to solve the above technical problems, embodiments of the present invention provide a merchant recommendation method, including: establishing a merchant recommendation model in advance, wherein the merchant recommendation model includes at least a conversion rate model and a return on investment model; The data determines the weights of the conversion rate model and the ROI model in the recommendation model; obtains the merchant's score according to the recommendation model; and performs merchant recommendation according to the merchant's score.

Embodiments of the present invention also provide a merchant recommendation device, comprising: a model establishment module for pre-establishing a merchant recommendation model, wherein the merchant recommendation model at least includes a conversion rate model and a return on investment model; the model establishes The module is also used to determine the weight of the conversion rate model and the ROI model in the recommendation model according to the user's historical behavior data; the score acquisition module is used to obtain the merchant's score according to the recommendation model; recommend The merchant module is used to recommend merchants according to the merchant ratings.

Embodiments of the present invention also provide an electronic device, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores a program executable by the at least one processor. instructions, the instructions are executed by the at least one processor, so that the at least one processor can execute: establish a merchant recommendation model in advance, wherein the merchant recommendation model includes at least a conversion rate model and a return on investment model; according to The user's historical behavior data determines the weight of the conversion rate model and the ROI model in the recommendation model; obtains the merchant's score according to the recommendation model; and recommends the merchant according to the merchant's score.

Embodiments of the present invention also provide a non-volatile storage medium for storing a computer-readable program, where the computer-readable program is used for a computer to execute the above merchant recommendation method.

Compared with the prior art, the embodiment of the present invention establishes a merchant recommendation model in advance, wherein the merchant recommendation model includes at least a conversion rate model and a return on investment model; the conversion rate model and all the investment return models are determined according to the user's historical behavior data. The weight of the ROI model in the recommendation model; the merchant's score is obtained according to the recommendation model; the merchant is recommended according to the merchant's score. Since the merchant recommendation model includes the conversion rate model and the ROI model, the merchants can be scored with multiple performance measurement indicators such as conversion rate and ROI; and the weights of the conversion rate model and ROI model are based on the user's history. The behavior data is determined, so that the recommendation model can give different scores for merchants according to the needs of different users, which helps to recommend suitable merchants for users and effectively improves the user experience; at the same time, recommending merchants to more likely to carry out the next It also helps to improve the merchant's performance measurement indicators and income profits.

In addition, determining the weights of the conversion rate model and the ROI model in the recommendation model according to the user's historical behavior data, specifically includes: determining the conversion rate model and the investment according to the user's historical behavior data The matching degree between the rate of return model and the user; according to the degree of matching between the conversion rate model and the rate of return on investment model and the user, determine the weight of the conversion rate model and the rate of return on investment model, so that the recommendation model is more realistic Effectively respond to user needs and help recommend more suitable merchants for users.

In addition, determining the matching degree of the conversion rate model and the ROI model with the user according to the user's historical behavior data includes: determining the conversion rate model and the user's historical behavior data according to the user's historical behavior data. The first matching degree of the user; according to the historical behavior data of the user, determine the second matching degree of the return on investment model and the user; according to the conversion rate model and the return on investment model and the user determining the weight of the conversion rate model and the ROI model, including: determining the weight of the conversion rate model according to the first matching degree; wherein, the higher the first matching degree, the The greater the weight of the conversion rate model; the weight of the ROI model is determined according to the second matching degree; wherein, the higher the second matching degree, the greater the weight of the ROI model. The higher the matching degree between the model and the user, the more accurate the model's estimation of the user's historical behavior data. Therefore, the model with higher matching degree has a higher weight, so that the merchant's score calculated according to the weighted recommendation model is higher. That is, it makes the merchants recommended to users more accurate and suitable, which helps to improve user experience.

In addition, determining the weight of the conversion rate model according to the first matching degree includes: obtaining, according to a preset first correspondence between the conversion rate matching degree and the conversion rate weight The weight corresponding to the first matching degree is used as the weight of the conversion rate model; the determining the weight of the ROI model according to the second matching degree includes: according to the preset ROI matching degree and The second correspondence of the ROI weights, and the weight corresponding to the second matching degree in the second correspondence is obtained as the weight of the ROI model.

In addition, determining the first degree of matching between the conversion rate model and the user according to the user's historical behavior data includes: obtaining, according to the user's historical behavior data, the conversion of the user by the conversion rate model according to the accuracy of the conversion rate prediction result; determining the first matching degree; determining the ROI model and the user's first matching degree according to the user's historical behavior data The second matching degree includes: obtaining the accuracy of the ROI prediction result of the user by the ROI model according to the historical behavior data of the user; determining the accuracy of the ROI prediction result based on the the second matching degree. The matching degree between the model and the user is obtained according to the prediction accuracy, which effectively eliminates the influence of user behavior deviation on the model, improves the accuracy and rationality of the matching degree between the obtained model and the user, and finally recommends a more suitable model for the user. merchant.

In addition, the merchant recommendation model also includes a click-through rate model, so as to take into account multiple effect measurement indicators of the merchant at the same time.

In addition, the merchant recommendation model is specifically:

score=modelctr*(1+alpha*nodelctcvr)*(1+beta*modelROI); wherein, the score represents the score of the merchant, the modelctr represents the click rate model, and the modelctcvr represents the conversion rate model, the alpha represents the weight of the conversion rate model, the modelROI represents the return on investment model, and the beta represents the weight of the return on investment model.

In addition, recommending merchants according to the merchant scores specifically includes: sorting the merchants to be recommended according to the high and low order of the scores of the merchants to be recommended; and recommending the merchants in the sorted order, so that the ranking The merchants in the forefront are merchants that better match the needs of users, which helps to improve the user experience, and also improves the performance measurement indicators and profit margins of the recommended merchants.

In addition, the user's historical behavior data includes one or any combination of the following data: the user's click behavior data, the user's order behavior data, and the user's order price data, so as to be closely related to the click rate and conversion rate of each merchant and performance measures such as ROI.

Description of drawings

1 is a flowchart of a merchant recommendation method according to a first embodiment of the present invention;

FIG. 2 is a flowchart of a method for determining model weights according to the first embodiment of the present invention;

3 is a flowchart of a method for determining model weights according to a second embodiment of the present invention;

4 is a structural diagram of a merchant recommendation device according to a third embodiment of the present invention;

FIG. 5 is a structural diagram of an electronic device according to a fourth embodiment of the present invention.

Detailed ways

In order to make the objectives, technical solutions and advantages of the embodiments of the present invention clearer, each embodiment of the present invention will be described in detail below with reference to the accompanying drawings. However, those of ordinary skill in the art will appreciate that, in the various embodiments of the present invention, many technical details are set forth in order for the reader to better understand the present application. However, even without these technical details and various changes and modifications based on the following embodiments, the technical solutions claimed in the present application can be realized. The following divisions of the various embodiments are for the convenience of description, and should not constitute any limitation on the specific implementation of the present invention, and the various embodiments may be combined with each other and referred to each other on the premise of not contradicting each other.

The first embodiment of the present invention relates to a merchant recommendation method, and the specific process is shown in FIG. 1 . In this embodiment, the merchant recommendation method in this embodiment is a dynamic process. Pre-establish a merchant recommendation model that includes at least a conversion rate model and an ROI model; whenever a recommendation request from a user is received, the merchant's score will be obtained according to the merchant recommendation model at the current moment. The weight of each model in the merchant recommendation model is determined by The user's historical behavior data is determined, and merchants are recommended based on merchant ratings. According to the rating of each merchant, real-time merchant recommendation is made to match user needs and improve user experience.

The following is a detailed description of the process in Figure 1:

Step 101: Establish a merchant recommendation model in advance.

Specifically, the merchant recommendation model includes at least a conversion rate model and an ROI model. For merchants, conversion rate and return on investment are both important indicators used to measure the effect of merchant recommendation; further, the merchant recommendation model also includes a click-through rate model, and the click-through rate is also an important indicator for users to measure the effect of merchant recommendation. The above conversion rate model, return on investment model and click rate model are all mature models that have been trained, and the historical recommendation data of merchants (such as the number of historical recommendations, historical clicks, historical orders, etc. of the merchant) Input to the above trained model, and the output of the model is the conversion rate, ROI and click rate of the merchant. The weights of the conversion rate model and the ROI model in the recommendation model are determined according to the user's historical behavior data, so that the recommendation model can give different scores for the merchants according to the needs of different users, which is helpful for Recommend suitable merchants for users, effectively improving the user experience. The pre-established merchant recommendation model in this embodiment organically combines the click-through rate model, the conversion rate model, and the ROI model, so as to take into account multiple effect measurement indicators of the merchant at the same time.

In step 102, a merchant rating is obtained according to the merchant recommendation model.

When obtaining the merchant's rating, input the merchant's historical recommendation data (such as the merchant's historical recommendation times, historical click times, historical order times, historical order price, etc.) into the merchant recommendation model, and the recommendation model can calculate the merchant's Click rate, conversion rate, and ROI; and calculate the merchant's score based on the corresponding weights of the conversion rate model and the ROI model. In this embodiment, a merchant rating is obtained for each merchant to be recommended, and finally a list of merchants matching the user's needs is recommended for the user.

Step 103, recommending a merchant according to the merchant's rating.

Specifically, the merchants to be recommended are sorted according to the order of their merchant ratings. Generally speaking, the higher the merchant rating of the merchant is, the more suitable the merchant is for the user, and the user is more likely to download the merchant. order, the merchant will be ranked first. In the sorted order, recommend businesses to users in the form of a recommended list.

In this embodiment, the weights of the conversion rate model and the ROI model in the recommendation model may be specifically determined through the process of FIG. 2 . The following is a detailed description of the process in Figure 2:

Step 201 , according to the historical behavior data of the user, determine the first matching degree between the conversion rate model and the user, and determine the second matching degree between the ROI model and the user.

Specifically, the user's historical behavior data includes one or any combination of the following data: the user's click behavior data, the user's order behavior data, and the user's order price data. Among them, the user's click behavior data reflects the merchants the user has clicked on, which is related to the click rate of each merchant; the user's order behavior data reflects the merchants that the user has actually consumed, which is related to the merchant's conversion rate and is used to determine the conversion rate model. The first matching degree with the user; the user's order price data reflects the actual consumption amount generated by the user at the merchant, that is, it reflects the income brought by the user to the merchant, which is related to the merchant's return on investment and is used to determine the return on investment model. The second match with the user.

Step 202: Determine the weight of the conversion rate model according to the first matching degree, and determine the weight of the ROI model according to the second matching degree.

Specifically, according to the preset first correspondence between the conversion rate matching degree and the conversion rate weight, the weight corresponding to the first matching degree in the first corresponding relationship is obtained as the weight of the conversion rate model; according to the preset investment return The second correspondence between the rate matching degree and the weight of the return on investment is obtained, and the weight corresponding to the second matching degree in the second correspondence is obtained as the weight of the return on investment model. For example, if the first matching degree is 70%, the weight corresponding to obtaining 70% matching degree in the first corresponding relationship is 0.7; if the second matching degree is 50%, then obtaining 50% matching degree in the second corresponding relationship corresponds to the weight. The corresponding weight is 0.5; in this embodiment, the higher the matching degree between the model and the user, the more accurate the model’s estimation of the user’s historical behavior data. Therefore, the model with higher matching degree has a larger weight, so that after determining the weight, The higher the merchant score calculated by the recommended recommendation model, the more accurate and suitable the merchant recommended to the user is, which helps to improve the user experience.

The merchant recommendation model in this embodiment can be expressed by the following formula:

score=modelctr*(1+alpha*modelctcvr)*(1+beta*modelROI)

Among them, score represents the merchant's score, modelctr represents the click rate model, modelctcvr represents the conversion rate model, alpha represents the weight of the conversion rate model, modelROI represents the return on investment model, and beta represents the weight of the ROI model; among them, alpha and beta initial It is a constant parameter obtained according to the test; in the process of merchant recommendation, the values of alpha and beta are adjusted according to the user's historical behavior data. For example, when the user's historical behavior data reflects that the merchant's conversion rate model is more closely matched with the user, the value of the weight alpha is increased.

The following is an example for specific description: for user A, according to the historical data of user A, it is determined that the weight of the conversion rate model is 0.8, and the weight of the ROI model is 0.5, then the alpha in the above formula is adjusted to 0.8, The beta is adjusted to 0.5, and the merchant recommendation model expressed by the weighted formula is used as the merchant recommendation model for User A; when a recommendation request from User A is received, the above-mentioned weight-adjusted formula is used to score the recommended merchants one by one , get the scores of each merchant to be recommended, for example, merchant A has a score of 90 points, merchant B has a score of 70 points, merchant C has a score of 86 points, merchant D has a score of 85 points, and merchant E has a score of 60 points. Then the merchants to be recommended are sorted as A-C-D-B-E and recommended to user A. Similarly, for user B, according to the historical data of user B, it is determined that the weight of the conversion rate model is 0.2, and the weight of the ROI model is 0.6, then the alpha in the above merchant recommendation model formula is adjusted to 0.2, and the beta is adjusted to 0.6, take the merchant recommendation model expressed by the formula after adjusting the weight as the merchant recommendation model for user B; when receiving a recommendation request from user B, use the above formula after adjusting the weight to score the recommended merchants one by one, and get each The ratings of the merchants to be recommended, for example, merchant A has a rating of 85 points, merchant D has a rating of 90 points, merchant E has a rating of 86 points, merchant F has a rating of 60 points, and merchant G has a rating of 70 points. The recommended merchants are ranked as D-E-A-G-F, and are recommended to user B.

Compared with the prior art, in this embodiment, a merchant recommendation model including at least a conversion rate model and an ROI model is established in advance, so as to simultaneously take into account multiple effect measurement indicators of the merchant; the conversion rate is determined according to the historical behavior data of the user. The matching degree of the model and the ROI model with the user respectively, and the weight of each model in the merchant recommendation model is determined according to each matching degree, so that the recommendation model can give different scores for the merchants according to the needs of different users; The model obtains merchant ratings; recommends merchants based on merchant ratings, so that the recommended merchants are suitable for users and match the user needs reflected in the user's historical behavior data, effectively improving the user experience; recommending merchants to users who are more likely to place orders, It is helpful to improve the performance measurement indicators such as the click rate, order rate, and return on investment of the recommended merchants, as well as the profit and profit of the merchants.

The second embodiment of the present invention relates to a merchant recommendation method. On the basis of the first embodiment, this embodiment further refines the method for determining the weights of the conversion rate model and the ROI model in the recommendation model. Specifically as shown in Figure 3.

Step 301 , according to the user's historical behavior data, obtain the accuracy of the user's conversion rate prediction result by the conversion rate model, and obtain the accuracy of the user's investment return rate prediction result by the ROI model.

Specifically, in this embodiment, the model evaluation AUC (Area Under roc Curve) index and the model evaluation GAUC (Group AUC) index are used to represent the accuracy of the prediction result. Calculate the AUC value for the conversion rate according to the historical order behavior data of each user; weight the calculated AUC value for the conversion rate, calculate the GAUC value for the conversion rate, and use the GAUC value for the conversion rate as the conversion rate model The accuracy of the user's conversion rate prediction results; similarly, according to the historical order price data of each user, calculate the AUC value for the return on investment; weight the calculated return on investment AUC value, calculate the return on investment The GAUC value of the rate of return, using the GAUC value for the rate of return on investment as the accuracy of the return on investment prediction result for the user by the rate of return on investment model. Through the above method of calculating the GAUC value, the influence of user behavior bias on the model is effectively eliminated, and the accuracy of the calculated prediction result is more reasonable.

More specifically, after the user's GAUC value is calculated, the user is classified, and the user is marked as a user belonging to the interval corresponding to the user's GAUC value; the interval corresponding to the GAUC value is set by itself as required. For example, if the interval corresponding to the GAUC value for the conversion rate is calculated as x, the conversion rate of the tagged user belongs to the x interval; if the interval corresponding to the GAUC value for the ROI is calculated as y, then the ROI of the tagged user belongs to y interval. Categorizing users in this way helps to subsequently determine the same matching degree and weight for users belonging to the same interval, and reduces the complexity and workload of the merchant recommendation process.

Step 302: Determine the first matching degree between the conversion rate model and the user according to the accuracy of the conversion rate prediction result; determine the second matching degree between the ROI model and the user according to the accuracy of the ROI prediction result.

Specifically, in this embodiment, the higher the calculated GAUC value, the higher the accuracy of the prediction result of the model, that is, the higher the matching degree between the model and the user. According to the preset relationship between the accuracy of the preset conversion rate prediction result and the first matching degree, the first matching degree between the conversion rate model and the user is determined; according to the preset accuracy of the ROI prediction result and the second matching degree The preset relationship determines the second degree of fit between the ROI model and the user.

Step 303: Determine the weight of the conversion rate model according to the first matching degree, and determine the weight of the ROI model according to the second matching degree. This step is substantially the same as step 202, and will not be repeated here. It should be noted that, after the weights of the conversion rate model and the ROI model are determined, the determined weights, corresponding to the interval to which the above-mentioned users belong, are used as the parameters for Users who belong to the same range can directly use the replaced formula to score the merchants to be recommended. For example, the replaced merchant recommendation model is represented by the following formula:

score _new =modelctr*(1+alpha ^x *modelctcvr)*(1+beta ^y *modelROI)

Among them, score _new represents the merchant's score, modelctr represents the click-through rate model, modelctcvr represents the conversion rate model, alpha ^x represents the weight of the replaced conversion rate model, and the superscript x represents the user's conversion rate in the x range; modelROI represents the return on investment. model, beta ^y represents the weight of the replaced ROI model, and the superscript y indicates that the user's ROI is in the y range.

Compared with the prior art, in this embodiment, the prediction accuracy of the model is obtained through the GAUC indicator according to the historical behavior data of the user, and the matching degree between the model and the user is obtained according to the prediction accuracy, which effectively eliminates the influence of user behavior deviation on the model. Influence, improve the accuracy and rationality of the matching degree between the obtained model and the user, and finally recommend more suitable merchants for the user. In addition, users are classified according to the GAUC indicator, which helps to reduce the complexity and workload of the merchant recommendation process.

The third embodiment of the present invention relates to a merchant recommendation device, as shown in FIG. 4 , including: a model establishment module 401 , a score acquisition module 402 and a recommended merchant module 403 .

The model establishment module 401 is used for pre-establishing a merchant recommendation model, wherein the merchant recommendation model at least includes a conversion rate model and a return on investment model.

The model building module 401 is further configured to determine the weights of the conversion rate model and the ROI model in the recommendation model according to the user's historical behavior data.

The score obtaining module 402 is configured to obtain the score of the merchant according to the recommendation model.

The recommending merchant module 403 is used for recommending merchants according to the merchant scores.

In one example, the model building module 401 is configured to determine the degree of matching of the conversion rate model and the ROI model with the user according to the historical behavior data of the user; according to the conversion rate model and the ROI model According to the matching degree with the user, the weights of the conversion rate model and the ROI model are determined.

In one example, the model building module 401 is configured to determine the first degree of matching between the conversion rate model and the user according to the historical behavior data of the user; and determine the return on investment according to the historical behavior data of the user The second matching degree between the rate model and the user; according to the first matching degree, the weight of the conversion rate model is determined; wherein, the higher the first matching degree, the greater the weight of the conversion rate model; The weight of the ROI model is determined according to the second matching degree; wherein, the higher the second matching degree is, the higher the weight of the ROI model is.

In one example, the model building module 401 is configured to obtain, according to a preset first correspondence between the conversion rate matching degree and the conversion rate weight, the weight corresponding to the first matching degree in the first correspondence, as The weight of the conversion rate model; according to the preset second correspondence between the matching degree of return on investment and the weight of return on investment, obtain the weight corresponding to the second matching degree in the second correspondence, as the weight of the second matching degree. The weights of the ROI model described above.

In one example, the model building module 401 is configured to obtain the accuracy of the conversion rate prediction result of the user by the conversion rate model according to the historical behavior data of the user; according to the accuracy of the conversion rate prediction result, Determine the first matching degree; obtain the accuracy of the ROI prediction result for the user by the ROI model according to the user's historical behavior data; The second degree of matching is determined.

In one example, the merchant recommendation model established by the model establishment module 401 further includes a click-through rate model.

In an example, the merchant recommendation model established by the model establishment module 401 is specifically:

score=modelctr*(1+alpha*modelctcvr)*(1+beta*modelROI)

The score represents the rating of the merchant, the modelctr represents the click rate model, the modelctcvr represents the conversion rate model, the alpha represents the weight of the conversion rate model, and the modelROI represents the ROI model, the beta represents the weight of the ROI model.

In one example, the recommending merchant module 403 is configured to sort the merchants to be recommended according to the high and low order of the scores of the merchants to be recommended, and recommend the merchants in the sorted order.

In an example, the historical behavior data of the user used by the score obtaining module 402 includes one or any combination of the following data: the user's click behavior data, the user's order behavior data, and the user's order price data.

It is not difficult to find that this embodiment is a device example corresponding to the first embodiment to the third embodiment, and this embodiment can be implemented in cooperation with the first embodiment to the third embodiment. The related technical details mentioned in the first embodiment to the third embodiment are still valid in this embodiment, and are not repeated here in order to reduce repetition. Correspondingly, the relevant technical details mentioned in this embodiment can also be applied to the first embodiment to the third embodiment.

It is worth mentioning that each module involved in this embodiment is a logical module. In practical applications, a logical unit may be a physical unit, a part of a physical unit, or multiple physical units. A composite implementation of the unit. In addition, in order to highlight the innovative part of the present invention, this embodiment does not introduce units that are not closely related to solving the technical problem proposed by the present invention, but this does not mean that there are no other units in this embodiment.

The fifth embodiment of the present invention relates to an electronic device. As shown in FIG. 5 , the electronic device includes:

at least one processor 501; and, a memory 502 communicatively connected to the at least one processor 501; and, a communication component 503 communicatively connected to the merchant recommendation device, the communication component 503 receiving and transmitting data under the control of the processor 501; wherein, The memory 502 stores instructions executable by the at least one processor 501, the instructions being executed by the at least one processor 501 to implement:

Establishing a merchant recommendation model in advance, wherein the merchant recommendation model includes at least a conversion rate model and a return on investment model;

Determine the weights of the conversion rate model and the ROI model in the recommendation model according to the user's historical behavior data;

Obtain the merchant's rating according to the recommendation model;

Merchant recommendation is performed according to the merchant rating.

Specifically, the terminal includes: one or more processors 501 and a memory 502, and one processor 501 is taken as an example in FIG. 4 . The processor 501 and the memory 502 may be connected through a bus or in other ways, and the connection through a bus is taken as an example in FIG. 4 . The memory 502, as a computer-readable storage medium, can be used to store computer software programs, computer-executable programs, and modules. The processor 501 executes various functional applications and data processing of the device by running the computer software programs, instructions and modules stored in the memory 502, ie, implements the above merchant recommendation method.

The memory 502 may include a storage program area and a storage data area, wherein the storage program area may store an operating system, an application program required by at least one function; the storage data area may store an option list and the like. Additionally, memory 502 may include high-speed random access memory, and may also include non-volatile memory, such as at least one magnetic disk storage device, flash memory device, or other non-volatile solid-state storage device. In some embodiments, the memory 502 may optionally include memory located remotely from the processor 501, and these remote memories may be connected to external devices via a network. Examples of such networks include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

One or more modules are stored in the memory 502, and when executed by the one or more processors 501, execute the merchant recommendation method in any of the above method embodiments.

The above product can execute the method provided by the embodiment of the present application, and has functional modules and beneficial effects corresponding to the execution method. For technical details not described in detail in this embodiment, please refer to the method provided by the embodiment of the present application.

In this embodiment, a merchant recommendation model is pre-established, wherein the merchant recommendation model includes at least a conversion rate model and a return on investment model; the conversion rate model and the return on investment model are determined according to the user's historical behavior data. The weight in the recommendation model; the rating of the merchant is obtained according to the recommendation model; the recommendation of the merchant is performed according to the rating of the merchant. In the embodiment of the present invention, the merchant recommendation model includes a conversion rate model and an ROI model, so that the merchant's score can take into account multiple effect measurement indicators such as conversion rate and ROI; and the weights of the conversion rate model and ROI model can be considered. , determined according to the user's historical behavior data, so that the recommendation model can give different scores for the merchants according to the needs of different users, which helps to recommend suitable merchants for users and effectively improves the user experience; Recommending it to users who are more likely to place an order will also help improve the merchant's performance measurement indicators and revenue.

The sixth embodiment of the present invention relates to a computer-readable storage medium storing a computer program. When the computer program is executed by the processor, the above embodiments of the merchant recommendation method are implemented.

That is, those skilled in the art can understand that all or part of the steps in the above embodiments of the merchant recommendation method can be completed by instructing the relevant hardware through a program, and the program is stored in a storage medium and includes several instructions to make a A device (which may be a single-chip microcomputer, a chip, etc.) or a processor (processor) executes all or part of the steps of the methods in the various embodiments of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, Read-Only Memory (ROM, Read-Only Memory), Random Access Memory (RAM, Random Access Memory), magnetic disk or optical disk and other media that can store program codes .

Those skilled in the art can understand that the above-mentioned embodiments are specific examples for realizing the present invention, and in practical applications, various changes in form and details can be made without departing from the spirit and the spirit of the present invention. scope.

The embodiments of the present application disclose A1. A merchant recommendation method, comprising:

Establishing a merchant recommendation model in advance, wherein the merchant recommendation model includes at least a conversion rate model and a return on investment model;

Determine the weights of the conversion rate model and the ROI model in the recommendation model according to the user's historical behavior data;

Obtain the merchant's rating according to the recommendation model;

Merchant recommendation is performed according to the merchant rating.

A2. The merchant recommendation method according to A1, wherein determining the weights of the conversion rate model and the ROI model in the recommendation model according to the user's historical behavior data, specifically including:

According to the historical behavior data of the user, determine the matching degree of the conversion rate model and the ROI model with the user;

According to the matching degree of the conversion rate model and the investment return model with the user, the weights of the conversion rate model and the investment return model are determined.

A3. The merchant recommendation method according to A2, wherein determining the degree of matching between the conversion rate model and the ROI model and the user according to the user's historical behavior data, comprising:

determining a first degree of matching between the conversion rate model and the user according to the user's historical behavior data;

determining a second degree of matching between the ROI model and the user according to the user's historical behavior data;

The determining the weights of the conversion rate model and the ROI model according to the degree of matching of the conversion rate model and the ROI model with the user, including:

Determine the weight of the conversion rate model according to the first matching degree; wherein, the higher the first matching degree, the greater the weight of the conversion rate model;

The weight of the ROI model is determined according to the second matching degree; wherein, the higher the second matching degree is, the higher the weight of the ROI model is.

A4. The merchant recommendation method according to A3, wherein determining the weight of the conversion rate model according to the first matching degree, comprising:

According to the preset first correspondence between the conversion rate matching degree and the conversion rate weight, obtain the weight corresponding to the first matching degree in the first corresponding relationship as the weight of the conversion rate model;

The determining the weight of the ROI model according to the second matching degree includes:

According to the preset second correspondence between the ROI matching degree and the ROI weight, the weight corresponding to the second matching degree in the second corresponding relationship is acquired as the weight of the ROI model.

A5. The merchant recommendation method according to A3, wherein determining the first degree of matching between the conversion rate model and the user according to the user's historical behavior data, comprising:

Acquiring the accuracy of the conversion rate prediction result of the user by the conversion rate model according to the historical behavior data of the user;

determining the first matching degree according to the accuracy of the conversion rate prediction result;

The determining of the second degree of matching between the ROI model and the user according to the user's historical behavior data includes:

According to the historical behavior data of the user, obtain the accuracy of the ROI prediction result of the user by the ROI model;

The second matching degree is determined according to the accuracy of the prediction result of the return on investment.

A6. The merchant recommendation method according to any one of A1 to A5, wherein the merchant recommendation model further includes a click-through rate model.

A7. The merchant recommendation method described in A6, the merchant recommendation model is specifically:

score=modelctr*(1+alpha*modelctcvr)*(1+beta*modelROI)

The score represents the rating of the merchant, the modelctr represents the click rate model, the modelctcvr represents the conversion rate model, the alpha represents the weight of the conversion rate model, and the modelROI represents the ROI model, the beta represents the weight of the ROI model.

A8. The merchant recommendation method according to A7, wherein the merchant recommendation is performed according to the merchant score, which specifically includes:

Sorting the merchants to be recommended according to the order of the scores of the merchants to be recommended;

Merchants are recommended in sorted order.

A9. The merchant recommendation method according to any one of A1 to A5, the historical behavior data of the user includes one of the following data or any combination thereof:

The user's click behavior data, the user's order behavior data, and the user's order price data.

The embodiment of the present application discloses B1. A merchant recommendation device, comprising:

a model establishment module for pre-establishing a merchant recommendation model, wherein the merchant recommendation model at least includes a conversion rate model and a return on investment model;

The model establishment module is further configured to determine the weights of the conversion rate model and the ROI model in the recommendation model according to the user's historical behavior data;

a score obtaining module, configured to obtain the merchant's score according to the recommendation model;

A merchant recommendation module is used to recommend merchants according to the merchant ratings.

The embodiment of the present application discloses C1. An electronic device, comprising:

at least one processor; and,

a memory communicatively coupled to the at least one processor; wherein,

The memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to execute:

Establishing a merchant recommendation model in advance, wherein the merchant recommendation model includes at least a conversion rate model and a return on investment model;

Determine the weights of the conversion rate model and the ROI model in the recommendation model according to the user's historical behavior data;

Obtain the merchant's rating according to the recommendation model;

Merchant recommendation is performed according to the merchant rating.

C2. The electronic device according to C1, wherein determining the weights of the conversion rate model and the ROI model in the recommendation model according to the user's historical behavior data, specifically including:

According to the historical behavior data of the user, determine the matching degree of the conversion rate model and the ROI model with the user;

According to the matching degree of the conversion rate model and the investment return model with the user, the weights of the conversion rate model and the investment return model are determined.

C3. The electronic device according to C2, wherein determining the degree of matching of the conversion rate model and the ROI model with the user according to the historical behavior data of the user, comprising:

determining a first degree of matching between the conversion rate model and the user according to the user's historical behavior data;

determining a second degree of matching between the ROI model and the user according to the user's historical behavior data;

The determining the weights of the conversion rate model and the ROI model according to the degree of matching of the conversion rate model and the ROI model with the user, including:

Determine the weight of the conversion rate model according to the first matching degree; wherein, the higher the first matching degree, the greater the weight of the conversion rate model;

The weight of the ROI model is determined according to the second matching degree; wherein, the higher the second matching degree is, the higher the weight of the ROI model is.

C4. The electronic device according to C3, wherein determining the weight of the conversion rate model according to the first matching degree, comprising:

According to the preset first correspondence between the conversion rate matching degree and the conversion rate weight, obtain the weight corresponding to the first matching degree in the first corresponding relationship as the weight of the conversion rate model;

The determining the weight of the ROI model according to the second matching degree includes:

According to the preset second correspondence between the ROI matching degree and the ROI weight, the weight corresponding to the second matching degree in the second corresponding relationship is acquired as the weight of the ROI model.

C5. The electronic device according to C3, wherein determining the first degree of matching between the conversion rate model and the user according to the user's historical behavior data, comprising:

Acquiring the accuracy of the conversion rate prediction result of the user by the conversion rate model according to the historical behavior data of the user;

determining the first matching degree according to the accuracy of the conversion rate prediction result;

The determining of the second degree of matching between the ROI model and the user according to the user's historical behavior data includes:

According to the historical behavior data of the user, obtain the accuracy of the ROI prediction result of the user by the ROI model;

The second matching degree is determined according to the accuracy of the prediction result of the return on investment.

C6. The electronic device according to any one of C1 to C5, wherein the merchant recommendation model further includes a click-through rate model.

C7. The electronic device according to C6, wherein the merchant recommendation model is specifically:

score=modelctr*(1+alpha*modelctcvr)*(1+beta*modelROI)

The score represents the rating of the merchant, the modelctr represents the click rate model, the modelctcvr represents the conversion rate model, the alpha represents the weight of the conversion rate model, and the modelROI represents the ROI model, the beta represents the weight of the ROI model.

C8. The electronic device according to C7, wherein the merchant recommendation is performed according to the merchant score, which specifically includes:

Sorting the merchants to be recommended according to the order of the scores of the merchants to be recommended;

Merchants are recommended in sorted order.

C9. The electronic device according to any one of C1 to C5, the historical behavior data of the user includes one of the following data or any combination thereof:

The user's click behavior data, the user's order behavior data, and the user's order price data.

The embodiment of the present application also discloses D1. A non-volatile storage medium for storing a computer-readable program, and the computer-readable program is used for a computer to execute the merchant recommendation as described in any one of A1 to A9 method.

Claims (10)

1. a merchant recommendation method, is characterized in that, comprises: Establishing a merchant recommendation model in advance, wherein the merchant recommendation model includes at least a conversion rate model and a return on investment model; Determine the weights of the conversion rate model and the ROI model in the recommendation model according to the user's historical behavior data; Obtain the merchant's rating according to the recommendation model; Merchant recommendation is performed according to the merchant rating. 2. The method for recommending merchants according to claim 1, wherein determining the weights of the conversion rate model and the ROI model in the recommendation model according to the user's historical behavior data, specifically comprising: According to the historical behavior data of the user, determine the matching degree of the conversion rate model and the ROI model with the user; According to the matching degree of the conversion rate model and the investment return model with the user, the weights of the conversion rate model and the investment return model are determined. 3. The method for recommending merchants according to claim 2, characterized in that, determining the matching degree of the conversion rate model and the ROI model with the user according to the historical behavior data of the user, comprising: determining a first degree of matching between the conversion rate model and the user according to the user's historical behavior data; determining a second degree of matching between the ROI model and the user according to the user's historical behavior data; The determining the weights of the conversion rate model and the ROI model according to the degree of matching of the conversion rate model and the ROI model with the user, including: Determine the weight of the conversion rate model according to the first matching degree; wherein, the higher the first matching degree, the greater the weight of the conversion rate model; The weight of the ROI model is determined according to the second matching degree; wherein, the higher the second matching degree is, the higher the weight of the ROI model is. 4. The merchant recommendation method according to claim 3, wherein the determining the weight of the conversion rate model according to the first matching degree comprises: According to the preset first correspondence between the conversion rate matching degree and the conversion rate weight, obtain the weight corresponding to the first matching degree in the first corresponding relationship as the weight of the conversion rate model; The determining the weight of the ROI model according to the second matching degree includes: According to the preset second correspondence between the ROI matching degree and the ROI weight, the weight corresponding to the second matching degree in the second corresponding relationship is acquired as the weight of the ROI model. 5. The method for recommending merchants according to claim 3, wherein determining the first degree of matching between the conversion rate model and the user according to the user's historical behavior data, comprising: Acquiring the accuracy of the conversion rate prediction result of the user by the conversion rate model according to the historical behavior data of the user; determining the first matching degree according to the accuracy of the conversion rate prediction result; The determining of the second degree of matching between the ROI model and the user according to the user's historical behavior data includes: According to the historical behavior data of the user, obtain the accuracy of the ROI prediction result of the user by the ROI model; The second matching degree is determined according to the accuracy of the prediction result of the return on investment. 6 . The merchant recommendation method according to claim 1 , wherein the merchant recommendation model further comprises a click-through rate model. 7 . 7. The merchant recommendation method according to claim 6, wherein the merchant recommendation model is specifically: score=modelctr*(1+alpha*modelctcvr)*(1+beta*modelROI) The score represents the rating of the merchant, the modelctr represents the click rate model, the modelctcvr represents the conversion rate model, the alpha represents the weight of the conversion rate model, and the modelROI represents the ROI model, the beta represents the weight of the ROI model. 8. A merchant recommendation device, comprising: a model establishment module for pre-establishing a merchant recommendation model, wherein the merchant recommendation model at least includes a conversion rate model and a return on investment model; The model establishment module is further configured to determine the weights of the conversion rate model and the ROI model in the recommendation model according to the user's historical behavior data; a score obtaining module, configured to obtain the merchant's score according to the recommendation model; A merchant recommendation module is used to recommend merchants according to the merchant ratings. 9. An electronic device, characterized in that, comprising: at least one processor; and, a memory communicatively coupled to the at least one processor; wherein, The memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to execute: Establishing a merchant recommendation model in advance, wherein the merchant recommendation model includes at least a conversion rate model and a return on investment model; Determine the weights of the conversion rate model and the ROI model in the recommendation model according to the user's historical behavior data; Obtain the merchant's rating according to the recommendation model; Merchant recommendation is performed according to the merchant rating. 10. A non-volatile storage medium for storing a computer-readable program, the computer-readable program being used for a computer to execute the merchant recommendation method according to any one of claims 1 to 7.