Exploration store data analysis method, system and equipment based on artificial intelligence and storage medium

Abstract

The invention discloses a shop exploration data analysis method, system and device based on artificial intelligence and a storage medium, and relates to the technical field of data analysis, and the method comprises the following steps: S1, building triple data, S2, screening shop nodes, S3, determining interest vectors, and S4, pushing shop information. According to the method, the catering data are obtained through the crawler tool and the triple is constructed, so that data structured storage can be realized, subsequent multi-dimensional statistical analysis and intelligent recommendation are facilitated, the utilization value of the data is improved, high-score shops corresponding to each user are screened out and stored in an interest set, and the user experience is improved. The user interest vector is generated according to the feature vector of the shop node in the interest set, the recommendation accuracy is improved, the potential preference of the user is mined, the user use vector is analyzed in combination with the user transaction record, and the interest vector and the use vector are combined to recommend the catering shop which the user does not go to by himself / herself. And the experience feeling of the user in the shop exploration process is further improved.

Description

Technical Field

[0001] The present invention relates to the field of data analysis technology, and specifically to an artificial intelligence-based store exploration data analysis method, system, device and storage medium. Background Art

[0002] The AI-based store exploration data analysis system is primarily an intelligent platform for analyzing massive amounts of store exploration data. The system can automatically analyze diverse data such as consumer ratings and consumer behavior, enabling consumers to obtain more personalized recommendations, enhance their consumer experience, and discover high-quality stores. Patent application number 202410690467.7 discloses a "store evaluation method, system, storage medium, and device based on store inspection data, relating to the field of data analysis technology. The method comprises: obtaining store inspection data for a target store; partitioning the store inspection data according to the indicator attributes of each evaluation indicator in a preset store evaluation rule to obtain sub-data corresponding to the evaluation indicator; extracting data content from the sub-data, the data content including image content and voice content; evaluating the evaluation indicator according to a preset template based on the data content to obtain first evaluation data corresponding to the evaluation indicator; and performing a comprehensive evaluation of the target store based on the first evaluation data corresponding to each evaluation indicator to obtain second evaluation data corresponding to the target store. This invention addresses the problem that existing store inspection evaluation methods are highly subjective, resulting in insufficient accuracy and hindering the continuous improvement and optimization of stores."

[0003] The above-mentioned existing technologies solve the problem that store inspection evaluations cannot timely and accurately guide target stores to make improvements and optimizations. However, during use, due to the unreasonable structured storage design of the acquired data, the efficiency of data retrieval is low. At the same time, it is also impossible to analyze the user's interests and preferences and transaction records during past store visits. Therefore, it is impossible to accurately predict the user's preference for catering stores that they have not visited in person, resulting in a poor user experience. Summary of the Invention

[0004] The purpose of the present invention is to provide an artificial intelligence-based store exploration data analysis method and system to solve the problems raised in the above background technology.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a store exploration data analysis method based on artificial intelligence, comprising the following steps:

[0006] S1. Construct triple data: Use a crawler tool to obtain catering platform data, store it in a database, count the total number of stores and construct multiple store nodes. After extracting the attribute variables of each store, construct corresponding attribute nodes based on the attribute variables of each store, and associate the store nodes with the attribute nodes to generate corresponding triple data;

[0007] S2. Filter store nodes: Construct multiple user nodes based on user data. Analyze the connection values between user nodes and store nodes using transaction records in the user data. After filtering out all non-zero connection values, query the corresponding store rating data based on the connection value and calculate the average rating. Store node triples with an average rating greater than eight are stored in the corresponding interest set.

[0008] S3. Determine the interest vector: Count the store node numbers with zero contact values, store them in the user node's candidate set, determine the feature vectors of all store nodes in the user node's interest set, and analyze the feature vectors in the interest set using a feature fusion algorithm to obtain the user node's interest vector.

[0009] S4. Push store information: After determining the usage sequence of the user node, transfer it to the GRU model for analysis to obtain the usage vector, aggregate the interest vector, usage vector and the low-dimensional representation of the store nodes in the selected set to obtain the preference coefficient of the store node, and transfer the store attribute variable value with a high preference coefficient to the user interface.

[0010] Preferably, said S1 comprises the following steps:

[0011] S101. Use a crawler tool to obtain restaurant store data and user data from an online restaurant platform, delete duplicate values, and store them in a database;

[0012] S102. After determining the total number of restaurants based on the restaurant data, construct multiple restaurant nodes using the total number of restaurants. After numbering each node, determine six attribute variables corresponding to each restaurant: restaurant name, cuisine, main dish, review, score, and address.

[0013] S103. After analyzing the restaurant name, cuisine, main dishes, comments, score, and address corresponding to each restaurant using the restaurant store data, construct multiple attribute nodes based on the restaurant name, cuisine, main dishes, comments, score, and address corresponding to each restaurant store;

[0014] S104. Construct association edges between the store node and each attribute node according to the type of attribute variables, generate corresponding triple data with the attribute variable names corresponding to the store node, attribute node and association edge, and combine all triple data to obtain a store attribute graph.

[0015] Preferably, said S2 comprises the following steps:

[0016] S201, construct multiple user nodes c1, c2, ..., c according to user data i,...,c n , and extract all store nodes s1,s2,…,s j ,…,s m , if the i-th user node c i There exists a store node s with the jth store node j Transaction records, then statistics c i and s j The number of transaction records between ij , according to N ij Determine c i and s j The connection value between ij , where I ij =N ij , m represents the total number of store nodes, n represents the total number of user nodes, i represents the user node number, j represents the store node number, if the i-th user node c i There is no store node s in the i-th store j Transaction records, then c i and s j The connection value between ij Set to zero;

[0017] S202: After counting all the contact values of the user node and the corresponding store node numbers, store them in the independent sequence set of the user node in sequence according to the transaction time in the transaction record.

[0018] Preferably, the step S2 further includes the following steps:

[0019] S203: Filter out all non-zero contact values in the independent sequence set of user nodes, analyze the corresponding store node numbers based on the contact values, query all rating data of the corresponding store in the user data according to the node numbers, and calculate the average store rating using the contact values and all rating data;

[0020] S204. After counting the average store ratings corresponding to the user nodes, a corresponding interest set is constructed for each user node. If the average store rating is greater than eight, the triple data corresponding to the store node in the store attribute graph is extracted according to the node number and stored in the interest set of the current user node. Otherwise, no operation is performed.

[0021] Preferably, the step S3 specifically includes the following steps:

[0022] S301: After counting all zero contact values in the independent sequence set of each user node, the corresponding store node number is analyzed according to the contact value, and the node number is stored in the candidate set of the user node;

[0023] S302. After determining the low-dimensional representation of the attribute node corresponding to each store node, select a store node in the interest set, use the node as the target node, and calculate the similarity between the target node and the low-dimensional representations of different attribute nodes in other store nodes in the interest set.

[0024] Preferably, the step S3 further includes the following steps:

[0025] S303: Determine the weight coefficients of each attribute variable corresponding to the target node based on the similarity, extract the low-dimensional representation of the attribute node of the target node, and calculate the feature vector of the current target node based on the weight coefficients and the low-dimensional representation of the attribute node;

[0026] S304: Repeat the operation until all the feature vectors of the store nodes in the interest set corresponding to the user node are analyzed, and then use the feature fusion analysis algorithm to analyze the feature vectors in the interest set corresponding to each user node to obtain the interest vector of the user node. The feature fusion analysis algorithm is specifically as follows:

[0027]

[0028] Among them, C γ represents the interest vector of the γth user node, α, β, γ represent parameters, g represents the number of store nodes in the interest set, s α represents the αth store node, h αβ Represents the low-dimensional representation of the βth attribute node of the αth store node, represents the interest set of the γth user node, ω αβ Represents the weight coefficient of the βth attribute variable of the αth store node.

[0029] Preferably, the S4 specifically includes the following steps:

[0030] S401. Construct a usage sequence of the user node according to the transaction time and store node number in the user node transaction record, and determine a low-dimensional representation of each store node according to the corresponding store node number in the usage sequence;

[0031] S402: After obtaining the low-dimensional representation of each store node in the usage sequence, the low-dimensional representation of each store node is sequentially transmitted to the GRU model for analysis according to the transaction time, and the usage vector of the user node is output;

[0032] S403: After counting the interest vector and usage vector of each user node, aggregate analysis is performed on them together with the low-dimensional representation of each store node in the candidate set, and the preference coefficients of different store nodes in the candidate set are calculated using an interest prediction algorithm;

[0033] S404: Filter out store nodes with preference coefficients higher than a preset value, and push attribute variable values corresponding to the store nodes to the current user interface.

[0034] The artificial intelligence-based store exploration data analysis system includes a store node determination unit, a user node construction unit, an interest vector analysis unit, and a preference coefficient generation unit;

[0035] The store node determination unit uses a crawler tool to obtain catering platform data, stores it in a database, counts the total number of stores and constructs multiple store nodes, extracts attribute variables of each store, constructs corresponding attribute nodes according to the attribute variables of each store, associates the store nodes with the attribute nodes, and thus generates corresponding triplet data;

[0036] The user node construction unit constructs multiple user nodes according to the user data, analyzes the connection value between the user node and the store node using the transaction records in the user data, filters out all the connection values that are not zero, queries the corresponding store rating data according to the connection value, calculates the average rating, and stores the store node triples with an average rating greater than eight in the corresponding interest set;

[0037] The interest vector analysis unit counts the store node numbers with zero contact values, stores them in a candidate set of the user node, determines the feature vectors of all store nodes in the user node interest set, and analyzes the feature vectors in the interest set using a feature fusion algorithm to obtain the interest vector of the user node;

[0038] After the preference coefficient generation unit determines the usage sequence of the user node, it transmits it to the GRU model for analysis to obtain the usage vector, and aggregates the interest vector, usage vector and low-dimensional representation of the store nodes in the selected set to obtain the preference coefficient of the store node, and transmits the store attribute variable value with a high preference coefficient to the user interface.

[0039] Compared with the prior art, the present invention has the following beneficial effects:

[0040] The present invention obtains catering data and constructs triples through crawler tools, which can realize structured data storage, facilitate subsequent multi-dimensional statistical analysis and intelligent recommendation, thereby improving the utilization value of the data, and screens out high-scoring stores corresponding to each user and stores them in the interest set, generates user interest vectors based on the feature vectors of store nodes in the interest set, improves recommendation accuracy, and thus explores user potential preferences, and at the same time analyzes user usage vectors in combination with user transaction records, combines interest vectors with usage vectors to recommend catering stores that users have not visited in person, and further enhances the user's experience in the store exploration process. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 Provides an overall method flow chart for an embodiment of the present invention;

[0042] Figure 2 A flow chart of a method for constructing triple data provided by an embodiment of the present invention;

[0043] Figure 3 A flow chart of a method for determining an interest vector provided by an embodiment of the present invention;

[0044] Figure 4 This is a flow chart of a method for pushing store information provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0045] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0046] See also Figures 1-4 The present invention provides a technical solution: a store exploration data analysis method based on artificial intelligence, comprising the following steps:

[0047] S1. Construct triple data: Use a crawler tool to obtain catering platform data, store it in a database, count the total number of stores and construct multiple store nodes. After extracting the attribute variables of each store, construct corresponding attribute nodes based on the attribute variables of each store, and associate the store nodes with the attribute nodes to generate corresponding triple data;

[0048] S2. Filter store nodes: Construct multiple user nodes based on user data. Analyze the connection values between user nodes and store nodes using transaction records in the user data. After filtering out all non-zero connection values, query the corresponding store rating data based on the connection value and calculate the average rating. Store node triples with an average rating greater than eight are stored in the corresponding interest set.

[0049] S3. Determine the interest vector: Count the store node numbers with zero contact values, store them in the user node's candidate set, determine the feature vectors of all store nodes in the user node's interest set, and analyze the feature vectors in the interest set using a feature fusion algorithm to obtain the user node's interest vector.

[0050] S4. Push store information: After determining the usage sequence of the user node, transfer it to the GRU model for analysis to obtain the usage vector, aggregate the interest vector, usage vector and the low-dimensional representation of the store nodes in the selected set to obtain the preference coefficient of the store node, and transfer the store attribute variable value with a high preference coefficient to the user interface.

[0051] Said S1 comprises the following steps:

[0052] S101. Use a crawler tool to obtain restaurant store data and user data from an online restaurant platform, delete duplicate values, and store them in a database;

[0053] S102. After determining the total number of restaurants based on the restaurant data, construct multiple restaurant nodes using the total number of restaurants. After numbering each node, determine six attribute variables corresponding to each restaurant: restaurant name, cuisine, main dish, review, score, and address.

[0054] S103. After analyzing the restaurant name, cuisine, main dishes, comments, score, and address corresponding to each restaurant using the restaurant store data, construct multiple attribute nodes based on the restaurant name, cuisine, main dishes, comments, score, and address corresponding to each restaurant store;

[0055] S104: Constructing association edges between the store node and each attribute node according to the attribute variable type, generating corresponding triples of data using the store node, attribute node, and attribute variable names corresponding to the association edge, and combining all triples of data to obtain a store attribute graph;

[0056] The S2 comprises the following steps:

[0057] S201, construct multiple user nodes c1, c2, ..., c according to user data i ,…,c n , and extract all store nodes s1,s2,…,s j ,…,s m , if the i-th user node c i There exists a store node s with the jth store node j Transaction records, then statistics c i and s j The number of transaction records between ij , according to N ij Determine c i and s j The connection value between ij , where I ij =N ij, m represents the total number of store nodes, n represents the total number of user nodes, i represents the user node number, j represents the store node number, if the i-th user node c i There is no store node s with the i-th store node j Transaction records, then c i and s j The connection value between ij Set to zero;

[0058] S202: After counting all the contact values of the user node and the corresponding store node numbers, they are stored in an independent sequence set of the user node in order of transaction time in the transaction record;

[0059] The S2 specifically further includes the following steps:

[0060] S203: Filter out all non-zero contact values in the independent sequence set of user nodes, analyze the corresponding store node numbers based on the contact values, query all rating data of the corresponding store in the user data according to the node numbers, and calculate the average store rating using the contact values and all rating data;

[0061] S204: After counting the average store ratings corresponding to the user node, construct a corresponding interest set for each user node. If the store rating average is greater than eight, extract the triple data corresponding to the store node in the store attribute graph according to the node number and store it in the interest set of the current user node. Otherwise, do nothing.

[0062] The S3 specifically includes the following steps:

[0063] S301: After counting all zero contact values in the independent sequence set of each user node, the corresponding store node number is analyzed according to the contact value, and the node number is stored in the candidate set of the user node;

[0064] S302: After determining the low-dimensional representation of the attribute node corresponding to each store node, select a store node in the interest set, use the node as the target node, and calculate the similarity between the target node and the low-dimensional representations of different attribute nodes in other store nodes in the interest set;

[0065] The S3 specifically further includes the following steps:

[0066] S303: Determine the weight coefficients of each attribute variable corresponding to the target node based on the similarity, extract the low-dimensional representation of the attribute node of the target node, and calculate the feature vector of the current target node based on the weight coefficients and the low-dimensional representation of the attribute node;

[0067] S304: Repeat the operation until all the feature vectors of the store nodes in the interest set corresponding to the user node are analyzed, and then use the feature fusion analysis algorithm to analyze the feature vectors in the interest set corresponding to each user node to obtain the interest vector of the user node. The feature fusion analysis algorithm is specifically as follows:

[0068]

[0069] Among them, C γ represents the interest vector of the γth user node, α, β, γ represent parameters, g represents the number of store nodes in the interest set, s α represents the αth store node, h αβ Represents the low-dimensional representation of the βth attribute node of the αth store node, represents the interest set of the γth user node, ω αβ represents the weight coefficient of the βth attribute variable of the αth store node;

[0070] The S4 specifically includes the following steps:

[0071] S401. Construct a usage sequence of the user node according to the transaction time and store node number in the user node transaction record, and determine a low-dimensional representation of each store node according to the corresponding store node number in the usage sequence;

[0072] S402: After obtaining the low-dimensional representation of each store node in the usage sequence, the low-dimensional representation of each store node is sequentially transmitted to the GRU model for analysis according to the transaction time, and the usage vector of the user node is output;

[0073] S403: After counting the interest vector and usage vector of each user node, perform aggregation analysis with the low-dimensional representation of each store node in the candidate set, and calculate the preference coefficients of different store nodes in the candidate set using an interest prediction algorithm. The interest prediction algorithm is specifically as follows:

[0074]

[0075] Among them, P xy represents the preference coefficient of the y-th store node corresponding to the x-th user node, σ(·) represents the sigmoid nonlinear function, represents the fusion result of the interest vector and usage vector of the x-th user node, represents the fusion result after transposition, s y represents the low-dimensional representation of the y-th store node, where x and y represent parameters;

[0076] S404: Filter out store nodes whose preference coefficients are higher than a preset value, and push attribute variable values corresponding to the store nodes to the current user interface;

[0077] An artificial intelligence-based store exploration data analysis system includes a store node determination unit, a user node construction unit, an interest vector analysis unit, and a preference coefficient generation unit;

[0078] The store node determination unit uses a crawler tool to obtain catering platform data, stores it in a database, counts the total number of stores and constructs multiple store nodes, extracts attribute variables of each store, constructs corresponding attribute nodes according to the attribute variables of each store, associates the store nodes with the attribute nodes, and thus generates corresponding triplet data;

[0079] The user node construction unit constructs multiple user nodes according to the user data, analyzes the connection value between the user node and the store node using the transaction records in the user data, filters out all the connection values that are not zero, queries the corresponding store rating data according to the connection value, calculates the average rating, and stores the store node triples with an average rating greater than eight in the corresponding interest set;

[0080] The interest vector analysis unit counts the store node numbers with zero contact values, stores them in a candidate set of the user node, determines the feature vectors of all store nodes in the user node interest set, and analyzes the feature vectors in the interest set using a feature fusion algorithm to obtain the interest vector of the user node;

[0081] After the preference coefficient generation unit determines the usage sequence of the user node, it transmits it to the GRU model for analysis to obtain the usage vector, and aggregates the interest vector, usage vector and low-dimensional representation of the store nodes in the selected set to obtain the preference coefficient of the store node, and transmits the store attribute variable value with a high preference coefficient to the user interface.

[0082] The present application also discloses an electronic device, which may include at least one processor, at least one memory, a power supply, a communication interface, an input / output interface, and a communication bus. The memory is used to store a computer program, which is loaded and executed by the processor to implement the relevant steps of the node management method disclosed in any of the aforementioned embodiments. Furthermore, the electronic device in this embodiment may be a computer.

[0083] In this embodiment, the power supply is used to provide operating voltage for each hardware device on the electronic device; the communication interface can create a data transmission channel between the electronic device and the external device. The communication protocol it follows is any communication protocol that can be applied to the technical solution of this application and is not specifically limited here; the input and output interface is used to obtain external input data or output data to the outside world. The specific interface type can be selected according to specific application needs and is not specifically limited here.

[0084] In addition, the memory as a carrier for resource storage can be a read-only memory, random access memory, disk or CD, etc. The resources stored thereon can include an operating system, computer programs, etc., and the storage method can be temporary storage or permanent storage.

[0085] The operating system is used to manage and control the hardware devices and computer programs on the electronic device, and can be Windows Server, Netware, Unix, Linux, etc. In addition to including a computer program capable of implementing the node management method executed by the electronic device disclosed in any of the aforementioned embodiments, the computer program can further include a computer program capable of implementing other specific tasks.

[0086] Furthermore, this application also discloses a computer-readable storage medium for storing a computer program; wherein, when executed by a processor, the computer program implements the aforementioned disclosed node management method. The specific steps of this method can be referred to the corresponding contents disclosed in the aforementioned embodiments and will not be repeated here.

[0087] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that includes a list of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus.

[0088] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A store exploration data analysis method based on artificial intelligence, characterized in that: The method comprises the following steps: S1. Construct triple data: Use a crawler tool to obtain catering platform data, store it in a database, count the total number of stores and construct multiple store nodes. After extracting the attribute variables of each store, construct corresponding attribute nodes based on the attribute variables of each store, and associate the store nodes with the attribute nodes to generate corresponding triple data; S2. Filter store nodes: Construct multiple user nodes based on user data. Analyze the connection values between user nodes and store nodes using transaction records in the user data. After filtering out all non-zero connection values, query the corresponding store rating data based on the connection value and calculate the average rating. Store node triples with an average rating greater than eight are stored in the corresponding interest set. S3. Determine the interest vector: Count the store node numbers with zero contact values, store them in the user node's candidate set, determine the feature vectors of all store nodes in the user node's interest set, and analyze the feature vectors in the interest set using a feature fusion algorithm to obtain the user node's interest vector. S4. Push store information: After determining the usage sequence of the user node, transfer it to the GRU model for analysis to obtain the usage vector, aggregate the interest vector, usage vector and the low-dimensional representation of the store nodes in the selected set to obtain the preference coefficient of the store node, and transfer the store attribute variable value with a high preference coefficient to the user interface.

2. The method for analyzing store exploration data based on artificial intelligence according to claim 1, characterized in that: Said S1 comprises the following steps: S101. Use a crawler tool to obtain restaurant store data and user data from an online restaurant platform, delete duplicate values, and store them in a database; S102. After determining the total number of restaurants based on the restaurant data, construct multiple restaurant nodes using the total number of restaurants. After numbering each node, determine six attribute variables corresponding to each restaurant: restaurant name, cuisine, main dish, review, score, and address. S103. After analyzing the restaurant name, cuisine, main dishes, comments, score, and address corresponding to each restaurant using the restaurant store data, construct multiple attribute nodes based on the restaurant name, cuisine, main dishes, comments, score, and address corresponding to each restaurant store; S104. Construct association edges between the store node and each attribute node according to the type of attribute variables, generate corresponding triple data with the attribute variable names corresponding to the store node, attribute node and association edge, and combine all triple data to obtain a store attribute graph.

3. The method for analyzing store exploration data based on artificial intelligence according to claim 1, characterized in that: The S2 comprises the following steps: S201, construct multiple user nodes c1, c2, ..., c according to user data i ,...,c n , and extract all store nodes s1,s2,...,s j ,...,s m , if the i-th user node c i There exists a store node s with the jth store node j Transaction records, then statistics c i and s j The number of transaction records between ij , according to N ij Determine c i and s j The connection value between ij , where I ij =N ij , m represents the total number of store nodes, n represents the total number of user nodes, i represents the user node number, j represents the store node number, if the i-th user node c i There is no store node s with the i-th store node j Transaction records, then c i and s j The connection value between ij Set to zero; S202: After counting all the contact values of the user node and the corresponding store node numbers, store them in the independent sequence set of the user node in sequence according to the transaction time in the transaction record.

4. The method for analyzing store exploration data based on artificial intelligence according to claim 3, characterized in that: The S2 specifically further includes the following steps: S203: Filter out all non-zero contact values in the independent sequence set of user nodes, analyze the corresponding store node numbers based on the contact values, query all rating data of the corresponding store in the user data according to the node numbers, and calculate the average store rating using the contact values and all rating data; S204. After counting the average store ratings corresponding to the user nodes, a corresponding interest set is constructed for each user node. If the average store rating is greater than eight, the triple data corresponding to the store node in the store attribute graph is extracted according to the node number and stored in the interest set of the current user node. Otherwise, no operation is performed.

5. The method for analyzing store exploration data based on artificial intelligence according to claim 1, characterized in that: The S3 specifically includes the following steps: S301: After counting all zero contact values in the independent sequence set of each user node, the corresponding store node number is analyzed according to the contact value, and the node number is stored in the candidate set of the user node; S302. After determining the low-dimensional representation of the attribute node corresponding to each store node, select a store node in the interest set, use the node as the target node, and calculate the similarity between the target node and the low-dimensional representations of different attribute nodes in other store nodes in the interest set.

6. The method for analyzing store exploration data based on artificial intelligence according to claim 5, characterized in that: The S3 specifically further includes the following steps: S303: Determine the weight coefficients of each attribute variable corresponding to the target node based on the similarity, extract the low-dimensional representation of the attribute node of the target node, and calculate the feature vector of the current target node based on the weight coefficients and the low-dimensional representation of the attribute node; S304. Repeat the operation until all the store node feature vectors in the interest set corresponding to the user node are analyzed, and then use the feature fusion analysis algorithm to analyze the feature vector in the interest set corresponding to each user node to obtain the interest vector of the user node.

7. The method for analyzing store exploration data based on artificial intelligence according to claim 1, characterized in that: The S4 specifically includes the following steps: S401. Construct a usage sequence of the user node according to the transaction time and store node number in the user node transaction record, and determine a low-dimensional representation of each store node according to the corresponding store node number in the usage sequence; S402: After obtaining the low-dimensional representation of each store node in the usage sequence, the low-dimensional representation of each store node is sequentially transmitted to the GRU model for analysis according to the transaction time, and the usage vector of the user node is output; S403: After counting the interest vector and usage vector of each user node, aggregate analysis is performed on them together with the low-dimensional representation of each store node in the candidate set, and the preference coefficients of different store nodes in the candidate set are calculated using an interest prediction algorithm; S404: Filter out store nodes with preference coefficients higher than a preset value, and push attribute variable values corresponding to the store nodes to the current user interface.

8. An artificial intelligence-based store exploration data analysis system, characterized in that: The store exploration data analysis system is applicable to the artificial intelligence-based store exploration data analysis method according to any one of claims 1 to 7, comprising a store node determination unit, a user node construction unit, an interest vector analysis unit, and a preference coefficient generation unit; The store node determination unit uses a crawler tool to obtain catering platform data, stores it in a database, counts the total number of stores and constructs multiple store nodes, extracts attribute variables of each store, constructs corresponding attribute nodes according to the attribute variables of each store, associates the store nodes with the attribute nodes, and thus generates corresponding triplet data; The user node construction unit constructs multiple user nodes according to the user data, analyzes the connection value between the user node and the store node using the transaction records in the user data, filters out all the connection values that are not zero, queries the corresponding store rating data according to the connection value, calculates the average rating, and stores the store node triples with an average rating greater than eight in the corresponding interest set; The interest vector analysis unit counts the store node numbers with zero contact values, stores them in a candidate set of the user node, determines the feature vectors of all store nodes in the user node interest set, and analyzes the feature vectors in the interest set using a feature fusion algorithm to obtain the interest vector of the user node; After the preference coefficient generation unit determines the usage sequence of the user node, it transmits it to the GRU model for analysis to obtain the usage vector, and aggregates the interest vector, usage vector and low-dimensional representation of the store nodes in the selected set to obtain the preference coefficient of the store node, and transmits the store attribute variable value with a high preference coefficient to the user interface.

9. An electronic device, characterized in that: The device comprises: a memory for storing a computer program; A processor, configured to implement the steps of the method according to any one of claims 1 to 7 when executing the computer program.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, which implements the steps of the method according to any one of claims 1 to 7 when executed.

Images