CN114385931A

CN114385931A - Method and device for obtaining recommendation form and electronic equipment

Info

Publication number: CN114385931A
Application number: CN202111647161.6A
Authority: CN
Inventors: 陈琳莉; 宁一鉴; 冯静; 刘冬寅; 张可人
Original assignee: China Telecom Corp Ltd
Current assignee: China Telecom Corp Ltd
Priority date: 2021-12-30
Filing date: 2021-12-30
Publication date: 2022-04-22

Abstract

A method, a device and an electronic device for obtaining a recommendation form are provided, wherein the method comprises the following steps: extracting category attribute keywords from input information sent by a user terminal, obtaining address information and a behavior portrait of the user terminal, screening all candidate merchants corresponding to the category attribute keywords from a candidate merchant recommendation list based on the address information, grading all candidate merchants based on the behavior portrait and commodity evaluation information corresponding to the candidate merchants, obtaining corresponding grading values of the candidate merchants respectively, generating a target recommendation form corresponding to the candidate merchants based on the grading values, and sending the target recommendation form to the user terminal. By the method, the target recommendation form which accords with the behavior portrait of the user side or the target recommendation form which is consistent with the category attribute keywords in the input information of the user side can be recommended to the user side in real time based on the position information of the user side.

Description

Method and device for obtaining recommendation form and electronic equipment

Technical Field

The application relates to the technical field of big data recommendation, in particular to a method and device for obtaining a recommendation form and electronic equipment.

Background

With the advent of the data economy era, in order to realize accurate matching between products of merchants and user requirements, a merchant recommendation algorithm for user requirements is introduced, the algorithm adopts a method that user tags of users and merchant tags of merchants are collected, the user tags record food type names, clothing type names and article names of the users searched at user terminals, the merchant tags record product type names and merchant names of the merchants, the names of the users searched at the user terminals are used as the user tags, when the user tags are consistent with the merchant tags, a server generates a merchant recommendation form based on all merchants corresponding to the merchant tags consistent with the user tags, and pushes the merchant recommendation form to the user terminals, so that accurate matching between the products of the merchants and the requirements of the users is realized.

Specifically, the process of pushing the merchant to the user side by the server using the merchant recommendation algorithm for the user requirement described above is as follows:

the name searched by the user A on the user end is the hot pot, after the server receives that the user label of the user is the hot pot, the server can match the commercial tenant of which the commercial tenant label is the hot pot, and the commercial tenant of which the commercial tenant label is the hot pot has the following characteristics: the method comprises the steps that merchants 1, merchants 2, merchants 3 and merchants 4 generate merchant recommendation forms after randomly arranging all matched merchants, and the merchant recommendation forms are pushed to a user end of a user A, when the merchant 1 is far away from the user end, and other users of the merchants 2 evaluate the merchant 2 as poor service, poor environment and low cost performance, the merchant 1 and the merchant 2 have high probability of being invalid merchants for the user A, but the merchant 1 and the merchant 2 still remain in the merchant recommendation forms, so that a large number of invalid merchants are doped in the merchant recommendation forms, and the accuracy of the merchant recommendation forms recommended to the user is low.

Disclosure of Invention

The method, the device and the electronic equipment for obtaining the recommended form sort merchants consistent with behavior portraits of the user side through address information of the user side and send the merchants to the user side, and sort merchants which belong to category attribute keywords in input information of the user side and recommend the merchants to the user side, so that successful butt joint of products of the merchants and requirements of the user side is achieved, and accuracy of the merchant form recommended to the user side is improved.

In a first aspect, the present application provides a method for obtaining a recommendation form, the method comprising:

extracting category attribute keywords from input information sent by a user terminal, and obtaining address information and a behavior portrait of the user terminal, wherein the behavior portrait is access information and access content generated by all merchants accessed by a user;

screening all candidate merchants corresponding to the category attribute keywords from a candidate merchant recommendation list based on the address information;

scoring all candidate merchants based on the behavior portrait and the commodity evaluation information corresponding to the candidate merchants to obtain scoring values corresponding to the candidate merchants respectively;

and generating a target recommendation form corresponding to the candidate merchant based on the credit value, and sending the target recommendation form to a user side.

In one possible design, before extracting the category attribute keyword from the input information sent by the user side, the method includes:

obtaining a first list, and converting the first list into a first merchant vector list based on a preset model, wherein the first list is a list generated based on the incidence relation between each user terminal and a merchant accessed by the user terminal;

obtaining a second merchant vector list generated according to merchants not accessed by a user terminal, and calculating similarity values between each first merchant vector in the first merchant vector list and each second merchant vector in the second merchant vector list;

generating a candidate merchant list based on all merchants corresponding to the similarity values within a preset range;

and calculating an access probability value corresponding to each candidate merchant in the candidate merchant list, and generating a candidate merchant recommendation list based on the candidate merchants corresponding to the arrangement sequence of the access probability values.

In one possible design, obtaining the address information of the user side includes:

acquiring the geographical position information of a WiFi access point corresponding to the user side, and taking the geographical position information of the WiFi access point as the address information of the user side; or

Acquiring an identity corresponding to a user side, and accessing a telecommunication operator network platform to which the user side belongs based on the identity to acquire address information of the user side.

In one possible design, obtaining a behavioral representation of the user end includes:

extracting basic information, network consumption information and keyword information in an input network of the user side from a network platform in communication connection with the user side;

and classifying the basic information, the network consumption information and the keyword information to generate a behavior portrait corresponding to the user side.

In one possible design, screening all candidate merchants corresponding to the category attribute keyword from a candidate merchant recommendation list based on the address information includes:

acquiring merchant address information corresponding to each candidate merchant in the candidate merchant recommendation list;

screening out a candidate merchant set corresponding to the address information of the user side within a preset distance from all the merchant address information;

and determining all candidate merchants corresponding to the category attribute keywords from the candidate merchant set according to the category attribute keywords.

In one possible design, scoring all candidate merchants based on the behavior representation and the commodity evaluation information corresponding to the candidate merchants to obtain respective corresponding scoring values of the candidate merchants includes:

extracting merchant names and merchant product information related to the candidate merchants in the behavior portrait, and evaluation types and evaluation description information in the commodity evaluation information;

calculating semantic similarity values among the merchant name, the merchant product information and the evaluation description information;

and performing weighted calculation based on the semantic similarity value and the evaluation type to obtain a score value corresponding to each candidate merchant.

In one possible design, generating a target recommendation form corresponding to the candidate merchant based on the scoring value includes:

arranging the scoring values according to a preset arrangement sequence to obtain all candidate commercial tenants arranged according to the arrangement sequence;

and taking all candidate merchants arranged according to the arrangement sequence as a target recommendation form corresponding to the candidate merchants.

In one possible design, obtaining a first list includes:

acquiring all user terminals and merchants visited by each user terminal;

deleting the user terminals with the number of 1 of the merchants visited within a first preset time and the merchants visited by the user terminals to obtain an initial list;

and deleting repeated merchants in the merchants visited by the user side in the initial list within second preset time to obtain the first list.

In one possible design, converting the first list into a first merchant vector list based on a preset model includes:

generating a first merchant sequence corresponding to merchants visited by each user terminal based on the time sequence corresponding to the merchants visited by each user terminal in the first list;

generating a merchant undirected graph corresponding to the first list based on a first merchant sequence corresponding to each user side;

recording a plurality of sequences randomly walked in the merchant undirected graph, inputting second merchant sequences corresponding to the sequences into a preset model, and obtaining a first merchant vector list corresponding to the second merchant sequences.

In one possible design, obtaining a second merchant vector list generated according to merchants not visited by a user side, and calculating a similarity value between each first merchant vector in the first merchant vector list and each second merchant vector in the second merchant vector list includes:

respectively calculating a similarity value between each first commercial tenant vector and each second commercial tenant vector;

and obtaining a similarity value between each first merchant vector in the first merchant vector list and each second merchant vector in the second merchant vector list.

In one possible design, generating a candidate merchant list based on all merchants corresponding to similarity values within a preset range includes:

screening all second merchant vectors corresponding to each first merchant vector in the first merchant vector list with the similarity value within a preset range;

sorting all second merchant vectors corresponding to the first merchant vector according to a preset sorting order of the similarity values, and screening out second merchants corresponding to the second merchant vectors with preset quantity values;

and obtaining all second merchants corresponding to all the first merchant vectors, and generating a candidate merchant list based on all the obtained second merchants.

In one possible design, calculating an access probability value corresponding to each candidate merchant in the candidate merchant list, and generating a candidate merchant recommendation list based on the candidate merchants corresponding to the ranking order of the access probability values includes:

obtaining a candidate merchant vector of each candidate merchant and extracting a merchant attention sequence corresponding to the merchant accessed by each user terminal from the first list according to the time of browsing the merchants of the user terminal;

inputting the merchant attention sequence of all the user terminals and the candidate merchant vector corresponding to the candidate merchant into a preset probability model to obtain an access probability value corresponding to each candidate merchant in the candidate merchant list;

and arranging the access probability values according to a preset arrangement sequence, and generating a candidate merchant recommendation list based on candidate merchants corresponding to the arrangement sequence.

In a second aspect, the present application provides an apparatus for obtaining a recommendation form, the apparatus comprising:

the acquisition module is used for extracting category attribute keywords from input information sent by a user side and acquiring address information and a behavior portrait of the user side;

the screening module is used for screening all candidate merchants corresponding to the category attribute keywords from the candidate merchant recommendation list based on the address information;

the evaluation module is used for grading all candidate merchants based on the behavior portrait and the commodity evaluation information corresponding to the candidate merchants to obtain the respective corresponding grading values of the candidate merchants;

and the recommending module is used for generating a target recommending form corresponding to the candidate commercial tenant based on the credit value and sending the target recommending form to the user side.

In a possible design, the obtaining module is specifically configured to obtain a first list, convert the first list into a first merchant vector list based on a preset model, obtain a second merchant vector list generated according to merchants not accessed by a user, calculate a similarity value between each first merchant vector in the first merchant vector list and each second merchant vector in the second merchant vector list, generate a candidate merchant list based on all merchants corresponding to the similarity value within a preset range, calculate an access probability value corresponding to each candidate merchant in the candidate merchant list, and generate a candidate merchant recommendation list based on the candidate merchants corresponding to the arrangement order of the access probability values.

In a possible design, the obtaining module is further configured to obtain all the user terminals and merchants visited by each user terminal, delete the user terminals with the number of 1 that access the merchants within a first preset time and the merchants visited by the user terminals, obtain an initial list, delete duplicate merchants in the merchants that the user terminals have accessed within a second preset time in the initial list, and obtain the first list.

In a possible design, the obtaining module is further configured to generate a first merchant sequence corresponding to the merchant to which each user terminal has accessed based on a time sequence corresponding to the merchant to which each user terminal has accessed in the first list, generate a merchant undirected graph corresponding to the first list based on the first merchant sequence corresponding to each user terminal, record a plurality of sequences randomly wandering in the merchant undirected graph, input a second merchant sequence corresponding to the plurality of sequences into a preset model, and obtain a first merchant vector list corresponding to the second merchant sequence.

In a possible design, the obtaining module is further configured to calculate a similarity value between each first merchant vector and each second merchant vector, and obtain a similarity value between each first merchant vector in the first merchant vector list and each second merchant vector in the second merchant vector list.

In a possible design, the obtaining module is further configured to screen out all second merchant vectors corresponding to each first merchant vector in the first merchant vector list with the similarity value within a preset range, sort all second merchant vectors corresponding to the first merchant vectors according to a preset arrangement order of the similarity value, screen out second merchants corresponding to second merchant vectors with preset quantity values, obtain all second merchants corresponding to all first merchant vectors, and generate the candidate merchant list based on all obtained second merchants.

In a possible design, the obtaining module is further configured to obtain a candidate merchant vector of each candidate merchant, extract a merchant attention sequence corresponding to a merchant to which each user side has accessed from the first list according to the time of browsing merchants of the user side, input the merchant attention sequences of all the user sides and the candidate merchant vector corresponding to the candidate merchant into a preset probability model, obtain an access probability value corresponding to each candidate merchant in the candidate merchant list, arrange the access probability values according to a preset arrangement order, and generate a candidate merchant recommendation list based on the candidate merchant corresponding to the arrangement order.

In a possible design, the obtaining module is further configured to obtain geographic location information of a WiFi access point corresponding to the user end, use the geographic location information of the WiFi access point as address information of the user end, or obtain an identity corresponding to the user end, and access a telecommunications operator network platform to which the user end belongs based on the identity to obtain the address information of the user end.

In a possible design, the obtaining module is further configured to extract basic information, network consumption information, and keyword information input into a network from a network platform to which the user is communicatively connected, classify the basic information, the network consumption information, and the keyword information, and generate a behavior representation corresponding to the user.

In a possible design, the screening module is specifically configured to obtain merchant address information corresponding to each candidate merchant in the candidate merchant recommendation list, screen out a candidate merchant set corresponding to the address information of the user terminal within a preset distance from all merchant address information, and determine all candidate merchants corresponding to the category attribute keyword from the candidate merchant set according to the category attribute keyword.

In a possible design, the scoring module is specifically configured to extract a merchant name and merchant product information related to the candidate merchant in the behavior sketch, and an evaluation type and evaluation description information in the commodity evaluation information, calculate a semantic similarity value between the merchant name and the evaluation description information, perform weighted calculation based on the semantic similarity value and the evaluation type, and obtain a score value corresponding to each candidate merchant.

In a possible design, the recommending module is specifically configured to arrange the score values according to a preset arrangement order, obtain all candidate merchants arranged according to the arrangement order, and use all candidate merchants arranged according to the arrangement order as a target recommending form corresponding to the candidate merchants.

In a third aspect, the present application provides an electronic device, comprising:

a memory for storing a computer program;

and the processor is used for realizing the steps of the method for obtaining the recommendation form when executing the computer program stored in the memory.

In a fourth aspect, a computer-readable storage medium has stored therein a computer program which, when executed by a processor, performs the above-mentioned method steps of obtaining a recommendation form.

For each of the first to fourth aspects and possible technical effects of each aspect, please refer to the above description of the possible technical effects for the first aspect or each possible solution in the first aspect, and no repeated description is given here.

Drawings

FIG. 1 is a flow chart of method steps for obtaining a recommendation form provided herein;

FIG. 2 is a schematic structural diagram of an apparatus for obtaining a recommendation form according to the present application;

fig. 3 is a schematic structural diagram of an electronic device provided in the present application.

Detailed Description

In order to make the objects, technical solutions and advantages of the present application more clear, the present application will be further described in detail with reference to the accompanying drawings. The particular methods of operation in the method embodiments may also be applied to apparatus embodiments or system embodiments. It should be noted that "a plurality" is understood as "at least two" in the description of the present application. "and/or" describes the association relationship of the associated objects, meaning that there may be three relationships, e.g., a and/or B, which may mean: a exists alone, A and B exist simultaneously, and B exists alone. A is connected with B and can represent: a and B are directly connected and A and B are connected through C. In addition, in the description of the present application, the terms "first," "second," and the like are used for descriptive purposes only and are not intended to indicate or imply relative importance nor order to be construed.

In the prior art, a user tag is generated based on a food type name, a clothing type name and an article name searched by a user terminal, so that the dimension singleness of the user tag of the user terminal is obtained, and in a merchant recommendation form generated by all merchants matched based on the user tag, the merchant position is far away from the position of the user terminal or the ordering of merchants in the merchant recommendation form is disordered, so that a large number of invalid merchants are mixed in the merchant recommendation form, and the accuracy of the merchant recommendation form recommended to the user is low.

In order to solve the above problem, the method for obtaining the recommendation form is adopted in the embodiments of the present application, and the method can recommend, to the user side, the merchant that conforms to the behavior representation of the user side and recommend the merchant level that is consistent with the category attribute keyword extracted from the input information of the user side based on the geographic position of the user side, and improve the accuracy of the merchant recommendation form recommended to the user side. The method and the device in the embodiment of the application are based on the same technical concept, and because the principles of the problems solved by the method and the device are similar, the device and the embodiment of the method can be mutually referred, and repeated parts are not repeated.

The embodiments of the present application will be described in detail below with reference to the accompanying drawings.

Referring to fig. 1, the present application provides a method for obtaining a recommendation form, which can improve the accuracy of a recommendation form recommended to a user side, and an implementation flow of the method is as follows:

step S1: extracting category attribute keywords from input information sent by a user side, and obtaining address information and a behavior portrait of the user side.

The embodiment of the application is used for recommending merchants consistent with the behavior portrait of the user side and merchants consistent with the category attribute keywords in the user side input information to the user side based on the address information of the user side, wherein the number of the merchants recommended to the user side is multiple, and all the merchants are sorted according to the scores, so that the user side can be enabled to quickly lock the optimal merchants.

It should be noted that, all merchants recommended to the user side are in the recommendation form, and the recommendation form is generated based on the candidate merchant recommendation list, in order to obtain the recommendation form, first, a candidate merchant recommendation list needs to be obtained, and a specific process of obtaining the candidate merchant recommendation list is as follows:

the method comprises the steps of obtaining all user terminals and merchants visited by each user terminal, screening all the user terminals in order to obtain an incidence relation between the user terminals and the merchants visited, detecting whether the number of the merchants visited by the user terminals in a first preset time is 1, and deleting the user terminals and the merchants with the number of the merchants 1 when the number of the merchants visited by the user in the first preset time is 1, so as to obtain an initial list corresponding to the user terminals.

After the initial list corresponding to the user terminal is obtained, in order to avoid that a large number of continuous repeat merchants are caused by repeatedly accessing one merchant within a second preset time from the merchants accessed by the user terminal, it is required to detect whether a repeat merchant exists in the second preset time from each merchant accessed by the user terminal in the initial list, and when it is determined that a repeat merchant exists within the second preset time, the repeat merchant is deleted, the visited merchant corresponding to each user terminal is traversed, and the above-described steps are repeated to obtain a first list corresponding to the initial list, where the first list records the association relationship between the user terminal and the merchant accessed by the user terminal.

After the first list is determined, in order to obtain the merchant sequence corresponding to the merchant to which each user terminal has accessed, the merchants to which the user terminal has accessed need to be arranged according to the time sequence of the merchants to which the user terminal has accessed to generate the merchant sequence corresponding to each user terminal, in order to better represent the relationship between merchants visited by the user terminal, a merchant undirected graph needs to be generated according to all the first merchant sequences, after the merchant undirected graph is obtained, a batch of second merchant sequences are generated based on a random walk mode, in the embodiment of the application, each merchant is converted into a vector and the incidence relation between the users and the merchants is shown, the decapwalk algorithm is adopted to learn the vector of each merchant, a first merchant vector list corresponding to a second merchant sequence is obtained, since the deepwalk algorithm is a technique well known to those skilled in the art, it will not be explained here too much.

After the first merchant list is obtained, since the merchants corresponding to the merchant vectors in the first merchant list are all merchants visited by the user terminal, in order to obtain other merchants similar to the merchants corresponding to the merchant vectors in the first merchant list, a second merchant vector list generated based on merchants not visited by the user terminal is obtained, and a manner of generating the second merchant vector list is the same as that of generating the first merchant vector list, which is not specifically described herein.

After the first merchant vector list and the second merchant vector list are obtained, in order to recall the second merchant vectors based on the first merchant vectors in the obtained first merchant vector list, similarity values between each first merchant vector and each second merchant vector are calculated, all the similarity values corresponding to the first merchant vectors are ranked, merchants related to the second merchant vectors corresponding to the similarity values within a preset range are screened out, and a candidate merchant list is generated based on all the screened merchants.

It should be noted that the similarity values described above may also be arranged from large to small, and the second merchant vector corresponding to the similarity value of the preset quantity value at the head of the arranged column is screened out, so that the merchant associated with the second merchant vector is determined based on the screened out second merchant vector, and the candidate merchant list is generated based on all screened out merchants.

After the candidate merchant list is obtained, since the merchants in the candidate merchant list are arranged randomly, in order to enable the user side to accurately lock the optimal user side from all merchant lists in the candidate merchant list, all merchants in the candidate merchant list need to be sorted, and in order to extract the merchant attention sequence of the user side, the merchants visited by the user side need to be analyzed according to the time of browsing merchants of the user side, and the analyzed result is used as the merchant attention sequence corresponding to the user side.

After the merchant attention sequence of the user side is obtained, a candidate merchant vector of each candidate merchant needs to be obtained, the merchant attention sequence of the user side and the candidate merchant vector corresponding to the candidate merchant are input into a preset probability model, and an access probability value corresponding to each candidate merchant in a candidate merchant list is obtained.

After the access probability value corresponding to each candidate merchant in the candidate merchant list is obtained, the access probability values are ranked according to the sequence from large to small, the candidate merchants corresponding to the access probability values are ranked according to the ranking sequence of the access probability values, and a candidate merchant recommendation list is generated based on all the ranked candidate merchants.

Through the method, the candidate merchant list is screened out from the mass merchant data, a mass of merchants which are not served by the user terminal or have low service frequency are screened out, all merchants in the candidate merchant list are guaranteed to be merchants with high user terminal access probability values, and the candidate merchants in the candidate merchant recommendation list are sorted according to the access probability values, so that the successful connection between the user terminal requirements and the products of the merchants is guaranteed.

After obtaining the candidate merchant recommendation list, when the user side does not input information, the target recommendation form can be pushed to the user side according to a preset period, when the user side is detected to have the input information, the category attribute keywords are extracted from the input information, and then the target recommendation form is pushed to the user side based on the extracted category attribute keywords.

After the category attribute keywords are extracted from the input information of the user side, in order to realize accurate recommendation for merchants with similar category attribute keywords, address information and behavior portraits of the user side need to be obtained, and the specific process of obtaining the address information of the user side is as follows:

the first method is as follows: and acquiring the geographical position information of the WiFi access point corresponding to the user terminal from the telecommunication network, acquiring the latest geographical position information of the WiFi access point of the user, and taking the geographical position information as the address information of the user terminal.

The second method comprises the following steps: the method comprises the steps of obtaining an identity of a user side, wherein the identity at least comprises identity information of a user to which the user side belongs and a contact telephone, accessing a telecommunication operator network platform to which the user side belongs based on the identity after obtaining the identity of the user side, and extracting address information of the user side from the user side information extracted based on the identity.

After the address information of the user terminal is obtained by adopting the above-described first or second mode, in order to obtain the behavior representation of the user terminal, the basic information and the network consumption information of the user terminal and the keyword information input into the network are extracted from the network platform in communication connection with the user terminal.

The basic information at least comprises identity information, health condition, contact information, work information, wealth information and family information, the network consumption information at least comprises communication behavior, customer experience, arrearage payment information, terminal information and sales information, the keyword information classification input into the network at least comprises information accessed by a user terminal, entertainment, communication, tools and online shopping, the information, the entertainment, the communication, the tools and the online shopping can be used as category attribute keywords, and each category attribute keyword can be further classified, such as: entertainment and entertainment can be further divided into music, games, videos, reading, cartoons, pictures and the like, other category attribute keywords can be further divided, the dividing standard can be classified according to the overall requirements of users, the main functions of applications and the contents of the applications, and the division of the other category attribute keywords refers to the further classification of the entertainment and is not described herein.

Generating a behavior portrait of the user terminal based on the basic information of the user terminal, the network consumption information and the input network keyword information, wherein the behavior portrait is access information and access content generated by all merchants accessed by the user.

Through the content described above, the candidate merchants in the candidate merchant recommendation list can be screened based on the address information of the user side, so that the probability of the user side accessing the candidate merchants is improved, and the consistency of products provided by the screened candidate merchants and the requirements of the user side can be ensured.

Step S2: and screening all candidate merchants corresponding to the category attribute keywords from the candidate merchant recommendation list based on the address information.

In the above description, the category attribute keyword of the user side is obtained, and since the extracted category attribute keyword is definitely in the candidate merchant classification name in the candidate merchant recommendation list, all candidate merchants corresponding to the classification name in the candidate merchant recommendation list consistent with the category attribute keyword can be determined based on the category attribute keyword.

Step S3: and scoring all candidate merchants based on the behavior portrait and the commodity evaluation information corresponding to the candidate merchants to obtain the scoring values respectively corresponding to the candidate merchants.

After obtaining all candidate merchants corresponding to the category attribute keyword, all merchants need to be ranked, and in order to achieve ranking of all candidate merchants, a candidate merchant image also needs to be obtained, which at least includes: basic details, operation attributes, visitor rate of the user side and commodity evaluation information of the candidate merchants.

The basic details of the candidate merchants include at least: the business hours of the candidate merchants, the per-capita consumption of the candidate merchants, the merchant addresses of the candidate merchants, the business circles to which the candidate merchants belong and the like; the operational attributes include at least: catering cate, leisure and recreation, clothes ornaments and the like; the visitor rate of the user side is the percentage of the number of the user sides consuming at the candidate commercial tenant divided by the total number of the user sides accessing the candidate commercial tenant, and when the visitor rate of the user side is higher, the candidate commercial tenant is more popular with the majority of users; the commodity evaluation information is divided into evaluation types and evaluation descriptions, and the evaluation types comprise the following types by taking the catering cate as an example: service, taste, cost/performance ratio, environment, etc., and the evaluation description includes: the descriptive contents of the candidate merchants of the user may be poor, good and excellent, and the evaluation types and evaluation descriptions of different types of candidate merchants are different, but it can be determined that each candidate merchant corresponds to one item evaluation information, and since the item evaluation information has diversity, it is not set forth herein too much.

After the candidate merchant image of each candidate merchant is obtained, the commodity evaluation information in each candidate merchant image is extracted, and then the candidate merchants are scored based on the behavior image of the user side obtained in the description and the commodity evaluation information to obtain the corresponding score values of the candidate merchants.

By the method, each candidate merchant is scored based on the behavior portrait of the user side and the commodity evaluation information of the candidate merchant, and the candidate merchants can be screened based on the scoring values of the candidate merchants or the candidate merchants to be consumed are determined based on the scoring values of the candidate merchants.

Step S4: and generating a target recommendation form corresponding to the candidate merchant based on the credit value, and sending the target recommendation form to a user side.

In the above description, the credit values corresponding to each candidate merchant are obtained, in order to ensure that the candidate merchants in the target recommendation form recommended to the user terminal are ranked, the credit values of all the candidate merchants need to be arranged in descending order, the ranked candidate merchants serve as the target recommendation form, and after the target recommendation form is obtained, the target recommendation form is sent to the user terminal.

By the method, the candidate merchant recommendation list is quickly screened out from a large number of merchants, the candidate merchants in the candidate merchant list are ranked according to the access probability value of the user terminal, appropriate candidate merchants are matched from the candidate merchant list through the behavior portrait of the user terminal and address position matching, and the screened candidate merchants are ranked according to the score value, so that the target recommendation form generated based on the ranked candidate merchants is ensured to be the optimal target recommendation form corresponding to the current time of the user terminal, and the accuracy of the target recommendation form pushed to the user terminal is improved.

Based on the same inventive concept, an embodiment of the present application further provides a device for obtaining a recommended form, where the device for obtaining a recommended form is used to implement a function of a method for obtaining a recommended form, and with reference to fig. 2, the device includes:

an obtaining module 201, configured to extract category attribute keywords from input information sent by a user, and obtain address information and a behavior portrait of the user;

a screening module 202, configured to screen all candidate merchants corresponding to the category attribute keyword from a candidate merchant recommendation list based on the address information;

the scoring module 203 is configured to score all candidate merchants based on the behavior representation and the commodity evaluation information corresponding to the candidate merchants to obtain respective corresponding score values of the candidate merchants;

and the recommending module 204 is configured to generate a target recommending form corresponding to the candidate merchant based on the credit value, and send the target recommending form to the user side.

In a possible design, the obtaining module 201 is specifically configured to obtain a first list, convert the first list into a first merchant vector list based on a preset model, obtain a second merchant vector list generated according to merchants not accessed by a user, calculate a similarity value between each first merchant vector in the first merchant vector list and each second merchant vector in the second merchant vector list, generate a candidate merchant list based on all merchants corresponding to the similarity values within a preset range, calculate an access probability value corresponding to each candidate merchant in the candidate merchant list, and generate a candidate merchant recommendation list based on the candidate merchants corresponding to the arrangement order of the access probability values.

In a possible design, the obtaining module 201 is further configured to obtain all the user terminals and merchants visited by each user terminal, delete the user terminals with the number of 1 that access the merchants within a first preset time and the merchants visited by the user terminals, obtain an initial list, delete duplicate merchants in the merchants that have been visited by the user terminals within a second preset time in the initial list, and obtain the first list.

In a possible design, the obtaining module 201 is further configured to generate a first merchant sequence corresponding to the merchant to which each user terminal has accessed based on a time sequence corresponding to the merchant to which each user terminal has accessed in the first list, generate a merchant undirected graph corresponding to the first list based on the first merchant sequence corresponding to each user terminal, record a plurality of sequences randomly wandering in the merchant undirected graph, input a second merchant sequence corresponding to the plurality of sequences into a preset model, and obtain a first merchant vector list corresponding to the second merchant sequence.

In a possible design, the obtaining module 201 is further configured to calculate a similarity value between each first merchant vector and each second merchant vector, and obtain a similarity value between each first merchant vector in the first merchant vector list and each second merchant vector in the second merchant vector list.

In a possible design, the obtaining module 201 is further configured to screen out all second merchant vectors corresponding to each first merchant vector in the first merchant vector list with the similarity value within a preset range, sort all second merchant vectors corresponding to the first merchant vectors according to a preset arrangement order of the similarity value, screen out second merchants corresponding to second merchant vectors with preset quantity values, obtain all second merchants corresponding to all first merchant vectors, and generate a candidate merchant list based on all obtained second merchants.

In a possible design, the obtaining module 201 is further configured to obtain a candidate merchant vector of each candidate merchant, extract a merchant attention sequence corresponding to a merchant to which each user terminal has accessed from the first list according to the time of browsing merchants of the user terminal, input the merchant attention sequences of all the user terminals and the candidate merchant vectors corresponding to the candidate merchants into a preset probability model, obtain an access probability value corresponding to each candidate merchant in the candidate merchant list, arrange the access probability values according to a preset arrangement order, and generate a candidate merchant recommendation list based on the candidate merchants corresponding to the arrangement order.

In a possible design, the obtaining module 201 is further configured to obtain geographic location information of a WiFi access point corresponding to the user end, use the geographic location information of the WiFi access point as address information of the user end, or obtain an identity corresponding to the user end, and access a telecommunications operator network platform to which the user end belongs based on the identity to obtain the address information of the user end.

In a possible design, the obtaining module 201 is further configured to extract basic information, network consumption information, and keyword information input into a network from a network platform to which the user is communicatively connected, classify the basic information, the network consumption information, and the keyword information, and generate a behavior representation corresponding to the user.

In a possible design, the screening module 202 is specifically configured to obtain merchant address information corresponding to each candidate merchant in the candidate merchant recommendation list, screen out a candidate merchant set corresponding to the address information of the user terminal within a preset distance from all merchant address information, and determine all candidate merchants corresponding to the category attribute keyword from the candidate merchant set according to the category attribute keyword.

In a possible design, the scoring module 203 is specifically configured to extract merchant names and merchant product information related to the candidate merchants in the behavior sketch, and evaluation types and evaluation description information in the commodity evaluation information, calculate semantic similarity values between the merchant names, the merchant product information, and the evaluation description information, perform weighted calculation based on the semantic similarity values and the evaluation types, and obtain a score value corresponding to each candidate merchant.

In a possible design, the recommending module 204 is specifically configured to arrange the score values according to a preset arrangement order, obtain all candidate merchants arranged according to the arrangement order, and use all candidate merchants arranged according to the arrangement order as a target recommending form corresponding to the candidate merchants.

Based on the same inventive concept, an embodiment of the present application further provides an electronic device, where the electronic device may implement the function of the foregoing apparatus for obtaining a recommended form, and with reference to fig. 3, the electronic device includes:

at least one processor 301 and a memory 302 connected to the at least one processor 301, in this embodiment, a specific connection medium between the processor 301 and the memory 302 is not limited in this application, and fig. 3 illustrates an example where the processor 301 and the memory 302 are connected through a bus 300. The bus 300 is shown in fig. 3 by a thick line, and the connection between other components is merely illustrative and not limited thereto. The bus 300 may be divided into an address bus, a data bus, a control bus, etc., and is shown with only one thick line in fig. 3 for ease of illustration, but does not represent only one bus or type of bus. Alternatively, the processor 301 may also be referred to as a controller, without limitation to name a few.

In the embodiment of the present application, the memory 302 stores instructions executable by the at least one processor 301, and the at least one processor 301 may execute the instructions stored in the memory 302 to perform a method for obtaining a recommendation form as discussed above. The processor 301 may implement the functions of the various modules in the apparatus shown in fig. 2.

The processor 301 is a control center of the apparatus, and may connect various parts of the entire control device by using various interfaces and lines, and perform various functions of the apparatus and process data by operating or executing instructions stored in the memory 302 and calling up data stored in the memory 302, thereby performing overall monitoring of the apparatus.

In one possible design, processor 301 may include one or more processing units, and processor 301 may integrate an application processor that primarily handles operating systems, user interfaces, application programs, and the like, and a modem processor that primarily handles wireless communications. It will be appreciated that the modem processor described above may not be integrated into the processor 301. In some embodiments, the processor 301 and the memory 302 may be implemented on the same chip, or in some embodiments, they may be implemented separately on separate chips.

The processor 301 may be a general-purpose processor, such as a Central Processing Unit (CPU), digital signal processor, application specific integrated circuit, field programmable gate array or other programmable logic device, discrete gate or transistor logic, discrete hardware components, or the like, that may implement or perform the methods, steps, and logic blocks disclosed in embodiments of the present application. A general purpose processor may be a microprocessor or any conventional processor or the like. The steps of the method for obtaining a recommendation form disclosed in the embodiments of the present application may be directly implemented by a hardware processor, or implemented by a combination of hardware and software modules in the processor.

Memory 302, which is a non-volatile computer-readable storage medium, may be used to store non-volatile software programs, non-volatile computer-executable programs, and modules. The Memory 302 may include at least one type of storage medium, and may include, for example, a flash Memory, a hard disk, a multimedia card, a card-type Memory, a Random Access Memory (RAM), a Static Random Access Memory (SRAM), a Programmable Read Only Memory (PROM), a Read Only Memory (ROM), a charge Erasable Programmable Read Only Memory (EEPROM), a magnetic Memory, a magnetic disk, an optical disk, and so on. The memory 302 is any other medium that can be used to carry or store desired program code in the form of instructions or data structures and that can be accessed by a computer, but is not limited to such. The memory 302 in the embodiments of the present application may also be circuitry or any other device capable of performing a storage function for storing program instructions and/or data.

The processor 301 is programmed to solidify the code corresponding to the method for obtaining a recommended form described in the foregoing embodiments into a chip, so that the chip can execute a step of obtaining a recommended form according to the embodiment shown in fig. 1 when running. How to program the processor 301 is well known to those skilled in the art and will not be described herein.

Based on the same inventive concept, the present application also provides a storage medium storing computer instructions, which when executed on a computer, cause the computer to execute the method for obtaining a recommendation form discussed above.

In some possible embodiments, the present application provides that the various aspects of a method of obtaining a recommendation form may also be implemented in the form of a program product comprising program code for causing a control apparatus to perform the steps of a method of obtaining a recommendation form according to various exemplary embodiments of the present application described above in this specification when the program product is run on a device.

As will be appreciated by one skilled in the art, embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.

The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to the application. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.

These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

It will be apparent to those skilled in the art that various changes and modifications may be made in the present application without departing from the spirit and scope of the application. Thus, if such modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is intended to include such modifications and variations as well.

Claims

1. A method of obtaining a recommendation form, comprising:

2. The method of claim 1, wherein before extracting the category attribute keyword from the input information sent from the user side, the method comprises:

3. The method of claim 1, wherein obtaining the address information of the user side comprises:

4. The method of claim 1, wherein obtaining the behavioral representation of the user terminal comprises:

5. The method of claim 1, wherein screening all candidate merchants from a candidate merchant recommendation list having the category attribute keyword based on the address information comprises:

6. The method of claim 1, wherein scoring all candidate merchants based on the behavior image and the product evaluation information corresponding to the candidate merchants to obtain respective scoring values corresponding to the candidate merchants comprises:

7. The method of claim 1, wherein generating the target recommendation form for the candidate merchant based on the scoring value comprises:

8. The method of claim 2, wherein obtaining the first list comprises:

acquiring all user terminals and merchants visited by each user terminal;

9. The method of claim 2, wherein converting the first list to a first merchant vector list based on a preset model comprises:

10. The method of claim 2, wherein obtaining a second merchant vector list generated according to merchants not visited by the user terminal, and calculating a similarity value between each first merchant vector in the first merchant vector list and each second merchant vector in the second merchant vector list comprises:

11. The method of claim 2, wherein generating the candidate merchant list based on all merchants corresponding to similarity values within a preset range comprises:

12. The method of claim 2, wherein calculating an access probability value corresponding to each candidate merchant in the candidate merchant list, and generating a candidate merchant recommendation list based on the candidate merchants corresponding to the ranking order of the access probability values comprises:

13. An apparatus for obtaining a recommendation form, comprising:

14. An apparatus, comprising:

a memory for storing a computer program;

a processor for implementing the method steps of any one of claims 1-12 when executing the computer program stored on the memory.

15. A computer-readable storage medium, characterized in that a computer program is stored in the computer-readable storage medium, which computer program, when being executed by a processor, carries out the method steps of any one of claims 1-12.