Point of interest recommendation method and apparatus based on association pattern and heterogeneous dual graph

Abstract

The application provides a point-of-interest recommendation method and device based on an association mode and a heterogeneous double graph, and a storage medium. The method comprises: obtaining a user information set, a point-of-interest information set, a point-of-interest category information set, and a user check-in time information set; constructing an association mode according to the user information set, the point-of-interest information set, the point-of-interest category information set, and the user check-in time information set; constructing a first heterogeneous graph and a second heterogeneous graph according to the association mode and the aforementioned sets; determining a first decomposition graph of the first heterogeneous graph and a second decomposition graph of the second heterogeneous graph based on vector embedding according to the association mode; determining a first information entropy and a first weight corresponding to the first heterogeneous graph and a second information entropy and a second weight corresponding to the second heterogeneous graph based on an information entropy weighting method; and determining a recommendation score of each point of interest according to the first information entropy, the first weight, the second information entropy, and the second weight. The application can improve the accuracy of point-of-interest recommendation.

Description

Technical Field

[0001] This document relates to the field of geographic information technology, and in particular to a method, apparatus and storage medium for recommending points of interest based on association patterns and heterogeneous dual maps. Background Technology

[0002] In recent years, point-of-interest (POI) recommendation systems based on location-based social networks (LBSNs) have significantly improved recommendation performance by integrating multi-dimensional contextual data such as user behavior, geographic information, time series, and social relationships.

[0003] Compared to traditional product or movie recommendation systems, POI recommendation faces more complex external dynamic factors and requires the analysis of heterogeneous interaction relationships between users and POIs. Current research uses graph embedding techniques to model user-POI interaction networks and, to some extent, captures the correlation of multimodal data.

[0004] However, existing POI recommendation systems still have the following key problems that lead to inaccurate recommendation results:

[0005] (1) Existing methods usually regard different contextual information as independent features of users and points of interest, such as time, space, semantics, etc., and fail to fully explore the inherent coupling relationship between users and points of interest based on features.

[0006] (2) Current node information aggregation methods based on path instances directly use the attention mechanism to allocate weights, which fails to quantify the differences in information contribution of different association patterns. Summary of the Invention

[0007] In view of the above solutions, this application aims to propose a method, apparatus and storage medium for recommending points of interest based on association patterns and heterogeneous dual graphs, so as to solve at least one of the above problems.

[0008] In a first aspect, one or more embodiments of this application provide a method for recommending points of interest based on association patterns and heterogeneous dual graphs, including:

[0009] Obtain a set of user information, a set of points of interest information, a set of points of interest category information, and a set of user check-in time information;

[0010] Based on the user information set, the point of interest information set, the point of interest category information set, and the user check-in time information set, at least one association pattern is constructed, which is used to represent the association relationship between user information, point of interest information, point of interest category information, and user check-in time information.

[0011] Based on the association pattern, the user information set, the interest point information set, the interest point category information set, and the user check-in time information set, a first heterogeneous graph of interest point information based on semantic relationships and a second heterogeneous graph of user information based on time association are constructed respectively.

[0012] Based on the association pattern, and using vector embedding, the first decomposition graph of the first heterogeneous graph and the second decomposition graph of the second heterogeneous graph are determined respectively.

[0013] Based on the information entropy weighting method, the first information entropy and the first weight corresponding to the first heterogeneous graph are determined;

[0014] Based on the information entropy weighting method, the second information entropy and the second weight corresponding to the second heterogeneous graph are determined.

[0015] The recommendation score for each point of interest is determined based on the first information entropy, the first weight, the second information entropy, and the second weight.

[0016] Furthermore, after obtaining the exploded graph, the method further includes:

[0017] Add negative samples to the first or second decomposition plot.

[0018] Furthermore, the inclusion of negative samples includes:

[0019] Set the total number of training rounds, the initial proportion of negative samples, and the maximum proportion of negative samples;

[0020] For each training round, the current negative sample ratio is adjusted based on the linear law and the initial negative sample ratio until the current negative sample ratio equals the highest negative sample ratio.

[0021] Furthermore, adjusting the current negative sample ratio based on the linear law and the initial negative sample ratio includes:

[0022] The current proportion of negative samples is calculated using the following formula:

[0023] ;

[0024] in, This represents the current proportion of negative samples. For the total number of training rounds, Indicates the current training round. For the initial ratio and It represents the highest proportion.

[0025] Furthermore, based on the information entropy weighting method, the first information entropy and the first weight corresponding to the first heterogeneous graph are determined, including:

[0026] Based on the information entropy weighting method, the information entropy of user-interest point interaction in the first decomposition graph is determined.

[0027] Based on the information entropy weighting method, the connection information entropy corresponding to the first decomposition graph is determined;

[0028] Based on the interaction information entropy and the connection information entropy, construct the information entropy matrix corresponding to the first decomposition graph;

[0029] Based on the information entropy matrix corresponding to the first decomposition graph, determine the first information entropy and the first weight corresponding to the first heterogeneous graph.

[0030] Further, based on the information entropy matrix corresponding to the first decomposition graph, the first weight corresponding to the first heterogeneous graph is determined, including:

[0031] The importance coefficient of the information entropy matrix is ​​calculated based on the attention mechanism;

[0032] The importance coefficient corresponding to the first heterogeneous graph is calculated based on the importance coefficient of the information entropy matrix.

[0033] Furthermore, based on the information entropy weighting method, the second information entropy and second weight corresponding to the second heterogeneous graph are determined, including:

[0034] Based on the information entropy weighting method, the information entropy of user-interest point interaction in the second decomposition graph is determined;

[0035] Based on the information entropy weighting method, the connection information entropy corresponding to the second decomposition graph is determined;

[0036] Based on the interaction information entropy and the connection information entropy, construct the information entropy matrix corresponding to the second decomposition graph;

[0037] Based on the information entropy matrix corresponding to the second decomposition graph, determine the second information entropy and the second weight corresponding to the second heterogeneous graph.

[0038] Furthermore, it is characterized in that,

[0039] Based on the information entropy matrix corresponding to the second decomposition graph, the second weights corresponding to the second heterogeneous graph are determined, including:

[0040] The importance coefficient of the information entropy matrix is ​​calculated based on the attention mechanism;

[0041] The importance coefficient of the second heterogeneous graph is calculated based on the importance coefficient of the information entropy matrix.

[0042] Secondly, one or more embodiments of this application provide an interest point recommendation device based on association patterns and heterogeneous dual graphs, including:

[0043] The acquisition module is used to acquire user information sets, interest point information sets, interest point category information sets, and user check-in time information sets.

[0044] The construction module is used to construct at least one association pattern based on the user information set, the interest point information set, the interest point category information set, and the user check-in time information set. The association pattern is used to represent the association relationship between user information, interest point information, interest point category information, and user check-in time information. Based on the user information set, the interest point information set, the interest point category information set, and the user check-in time information set, the module also constructs a first heterogeneous graph of interest point information based on semantic relationships and a second heterogeneous graph of user information based on time associations.

[0045] The data processing module is used to determine, based on the association pattern and vector embedding, a first decomposition graph of the first heterogeneous graph and a second decomposition graph of the second heterogeneous graph, respectively; to determine the first information entropy and first weight corresponding to the first heterogeneous graph based on the information entropy weighting method; to determine the second information entropy and second weight corresponding to the second heterogeneous graph based on the information entropy weighting method; and to determine the recommendation score of each interest point based on the first information entropy, the first weight, the second information entropy, and the second weight.

[0046] Thirdly, embodiments of this application provide a storage medium for storing computer-executable instructions, characterized in that, when executed, the computer-executable instructions implement the steps of the interest point recommendation method based on association patterns and heterogeneous dual graphs as described in the first aspect.

[0047] Compared with the prior art, this application can achieve at least the following technical effects:

[0048] This application first constructs a heterogeneous dual graph of interest points and users, laying the foundation for subsequent fusion of multiple factors. Second, it constructs association patterns and transforms the heterogeneous graph into a decomposed graph based on these patterns, enabling parameter fusion from multiple dimensions. Third, it determines the weights of the heterogeneous graph based on the decomposed graph, vector embedding, and information entropy weighting. Finally, based on the weights of the heterogeneous graph, it determines the recommendation score for each interest point. Therefore, this application can fully explore the feature-based intrinsic coupling relationship between users and interest points to improve the accuracy of recommendation results. Attached Figure Description

[0049] To more clearly illustrate the technical solutions in one or more embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0050] Figure 1 A flowchart illustrating an interest point recommendation method based on association patterns and heterogeneous dual graphs, provided for one or more embodiments of this application;

[0051] Figure 2 This is a schematic diagram of the structure of an interest point recommendation device based on association patterns and heterogeneous dual graphs, provided for one or more embodiments of this application. Detailed Implementation

[0052] To enable those skilled in the art to better understand the technical solutions in one or more embodiments of this application, the technical solutions in one or more embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of the embodiments. Based on one or more embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort should fall within the protection scope of this document.

[0053] This application provides a method for recommending points of interest based on association patterns and heterogeneous dual graphs, such as... Figure 1 As shown, it includes the following steps:

[0054] Step 1: Obtain the user information set, the point of interest information set, the point of interest category information set, and the user check-in time information set.

[0055] In this embodiment of the application, the user set is defined as The set of points of interest is Collection of interest point categories The set of user check-in times is ,in, , and These represent the number of users, the number of interest points, and the number of categories, respectively. This indicates the number of time periods divided, with one day divided into 24 time periods.

[0056] Step 2: Based on the user information set, the point of interest information set, the point of interest category information set, and the user check-in time information set, construct at least one association pattern.

[0057] In this embodiment, the association pattern is used to represent the relationship between user information, point-of-interest information, point-of-interest category information, and user check-in time information. Specifically, the association pattern is defined as:

[0058] If node and By belonging to type If other nodes are connected, then and There are indirect association patterns. ;like and By belonging to type If the edges are connected, then and There is a direct association pattern For example, relationships ,node and There is a time-related pattern; ,node and There is a direct sign-in association mode.

[0059] The association pattern is as follows:

[0060] Direct sign-in association mode and This represents the interaction information between users and points of interest, where Check-in indicates that the user is checking in at the point of interest; indirect association mode. This indicates that users connect through shared points of interest, capturing similarities or social preferences among users; indirect association pattern. This indicates that points of interest are connected through shared users, reflecting to some extent the potential geographical proximity between them; indirect association mode. This indicates that connections are established through category information of points of interest, reflecting semantic and functional similarity; indirect association mode. This indicates that users establish connections through shared access times, capturing the similarity in behavioral patterns among users.

[0061] Step 3: Based on the association pattern, user information set, interest point information set, interest point category information set, and user check-in time information set, construct a first heterogeneous graph of interest point information based on semantic relationships and a second heterogeneous graph of user information based on time association.

[0062] In this embodiment of the application, the first heterogeneous graph construction process is as follows:

[0063] Constructing heterogeneous graphs of interest points based on semantic relationships Wherein, the connection edge between the user and the point of interest is represented as The edges connecting points of interest to their respective categories are represented as follows: In the original dataset, points of interest only have name attributes. To obtain the interest band category nodes in the heterogeneous graph, this application uses a pre-trained word embedding model to calculate word vectors for all interest point names. Then, the K-means algorithm is used to cluster the word vectors. Based on the classification and grading standards for urban infrastructure construction, all interest point categories are unsupervisedly classified into 14 clusters to obtain the high-order prototype vector of each cluster. The calculation formula is:

[0064] ;

[0065] in These are the sample word vectors and their respective category labels. For the point of interest The prototype vectors of each category, For category Supported sample set To embed a neural network, we first utilize a direct sign-in association pattern. Node extraction from the semantic relationship graph yields an interest-user interaction decomposition graph centered on interest points. This is used to mine explicit interaction information between users and points of interest, setting the edge weights to the average normalized value of user check-in counts; then, the indirect association pattern is used in the same way. and Obtain the isomorphic decomposition graph of interest points and This method is used to uncover implicit matching relationships between points of interest, setting edge weights to the normalized average of the local distances between these points. Other association patterns can also be obtained using the same approach to create heterogeneous graphs.

[0066] The second heterogeneous graph construction process is as follows:

[0067] Constructing a heterogeneous graph of users based on time association The connection edges between points of interest and users are represented as follows: The connection edge between the user and the check-in time is represented as In this embodiment of the invention, a day is divided into 24-hour periods, and a unique ID value is assigned to each period for one-hot encoding as a feature of the time node. Then, an embedding method is used to map the ID value of the time node to an implicit vector space to obtain an initial vector representation. The calculation formula is as follows: ,in For each time node, a one-hot vector. This is the weight transformation matrix. This is a time-based feature vector. First, the direct sign-in association mode is utilized. Node extraction from the time-related graph yields a user-interest interaction decomposition graph centered on the user. This is used to mine explicit interaction information between users and points of interest, setting the edge weights to the average normalized value of check-in counts; then, the indirect association pattern is used in the same way. and Obtain the user isomorphic decomposition graph and This method is used to uncover implicit user relationships; other relationship patterns can also be obtained using the same approach to create heterogeneous graphs. The edge weight calculation formula is:

[0068] ;

[0069] in, and They are respectively and The neighborhood group, For logarithmic operations, for The degree, if and If two people visit points of common interest, they are considered neighbors.

[0070] Step 4: Based on the association pattern and vector embedding, determine the first decomposition graph of the first heterogeneous graph and the second decomposition graph of the second heterogeneous graph.

[0071] In this embodiment of the application, the method for determining the first exploded view and the second exploded view is specifically as follows:

[0072] Since users check in across multiple time periods, simple vector summation may not be able to learn the non-linear characteristics of users' time preferences. Given a user or points of interest Nodes are used to capture patterns in user behavior over time. The number of sign-ins in each time period is normalized, and the normalized value is used as a weight to weight each time feature vector.

[0073] For example, if a user checked in 10 times across the entire dataset—twice in the morning, six times at noon, and twice in the evening—the user might prefer to be active around noon. Therefore, the weights for these times would be 0.2 for morning, 0.6 for noon, and 0.2 for evening. Each time period has a time feature vector, which is multiplied by its respective weight and then summed to obtain a fused time feature vector. This method reduces the number of input vectors in subsequent calculations, thereby reducing computational overhead.

[0074] The weighted time feature vector and the user random embedding vector are then concatenated and input into a two-layer MLP to obtain the time-aware user feature vector.

[0075] The category attribute of interest points is one of the bases for model decisions. To increase the interpretability of the model, interest points are... The random embedding vector and the category embedding vector are input into a two-layer MLP to obtain category-aware interest point embeddings, so as to recommend certain interest points according to the user's category preferences.

[0076] Association Pattern The next node It forms a decomposition graph based on a specific pattern with its neighboring nodes, and then adds a certain proportion of these nodes to the decomposition graph. Join from The difficult negative samples selected from the unsigned nodes are used to obtain a decomposition graph with new samples. All node features are represented as The edge weights of negative samples are all set to -1. Then, all edge weights are normalized again. The purpose of mean normalization is to retain the negative weight information of negative samples and improve the model's ability to distinguish samples.

[0077] To obtain the feature embedding vectors of users and interest point nodes under the current association pattern, a graph attention network is used to aggregate node neighborhood information and incorporate their own features during computation. For a given pair of nodes... The formula for calculating the attention coefficient is:

[0078] ;

[0079] in, Represents a node Compared to Attention coefficient and These are parameters for shared learning. It is an activation function. It is a vector concatenation operation.

[0080] Then, the attention coefficient is normalized using the softmax function, calculated as follows:

[0081] ;

[0082] Last updated node features:

[0083] ;

[0084] in, express Features that incorporate neighborhood information It is an activation function.

[0085] Step 5: Based on the first decomposition graph, determine the first information entropy and the first weight corresponding to the first heterogeneous graph using the information entropy weighting method.

[0086] In this embodiment, multiple embedded representations of users and points of interest can be obtained according to different association pattern types. Therefore, the final user and point of interest node embeddings require further fusion. Users and points of interest can be obtained under all association patterns. _ embedding vectors, denoted as Use these vectors to construct a user embedding matrix: ,

[0087] Use these vectors to construct an interest point embedding matrix: , For the embedded dimension.

[0088] Then, based on these user embedding matrices and interest point embedding matrices, the importance of preference feature embeddings under different association patterns is evaluated using the information entropy weighting method. Afterwards, the weights of the heterogeneous graphs are determined based on these importance values.

[0089] Step 6: Based on the second decomposition graph, determine the second information entropy and the second weight corresponding to the second heterogeneous graph using the information entropy weighting method.

[0090] In this embodiment of the application, the method for determining the second information entropy and the second weight is the same as the method for determining the first information entropy and the first weight.

[0091] Step 7: Determine the recommendation score for each point of interest based on the first information entropy, the first weight, the second information entropy, and the second weight.

[0092] In this embodiment of the application, the specific process for determining the recommendation score is as follows:

[0093] The inner product of the end-user embedding vector and the interest point embedding vector is obtained. The recommendation score for each point of interest, as determined by the user, is calculated using the following formula:

[0094] ;

[0095] Then, the probability distribution of user access points of interest is obtained through MLP. Optimization function selection: root mean square error loss The calculation formula is:

[0096] ;

[0097] in For the true value, This indicates that the user has visited this point of interest. This indicates that the site has not been visited. Indicates that the user accessed the first The predicted probability of each point of interest.

[0098] In this embodiment, using only positive samples from the sign-in data during training ignores the implicit negative feedback information of unsigned interest points to users. Adding negative samples can better learn user interest and preference features. However, the ratio of positive to negative samples significantly affects model performance and recommendation effectiveness. Some methods often use a fixed proportion of positive samples in all samples and randomly select negative samples, which may lead to problems such as model overfitting, insufficient generalization ability, and sample trade-offs. Therefore, this application designs a method to linearly and adaptively adjust the composition of training samples based on evaluation metrics. Specifically, it sets the total number of training rounds, the initial proportion of negative samples, and the maximum proportion of negative samples.

[0099] For each training round, the current negative sample ratio is adjusted based on the linear law and the initial negative sample ratio until the current negative sample ratio equals the highest negative sample ratio.

[0100] The current proportion of negative samples is calculated using the following formula:

[0101] ;

[0102] in, This represents the current proportion of negative samples. For the total number of training rounds, Indicates the current training round. For the initial ratio and It represents the highest proportion.

[0103] Since there are other nodes in the negative samples that have different relevance to the current node, in order to enhance the model's ability to distinguish samples, cosine similarity is used to calculate the most similar difficult negative samples to the current node's features and add them to the decomposition graph. The formula for calculating cosine similarity is:

[0104] ;

[0105] in This represents the modulo operation of the node feature vectors. The rate of change of the training loss can significantly reflect the model's convergence performance, so the loss value is chosen as the evaluation criterion for adjusting the sample composition. If the average loss calculated from the wheel loss value is not greater than the threshold, then... If the loss value decreases slowly, it is considered difficult to adjust the proportion of negative samples.

[0106] In this embodiment of the application, the first information entropy and the first weight corresponding to the first heterogeneous graph are determined based on the information entropy weighting method. The specific process is as follows:

[0107] Based on the information entropy weighting method, the interaction information entropy between users and points of interest in the first decomposition graph is determined; based on the information entropy weighting method, the connection information entropy corresponding to the first decomposition graph is determined; based on the interaction information entropy and the connection information entropy, the information entropy matrix corresponding to the first decomposition graph is constructed; based on the information entropy matrix corresponding to the first decomposition graph, the first information entropy and the first weight corresponding to the first heterogeneous graph are determined.

[0108] In this embodiment of the application, the first weight corresponding to the first heterogeneous graph is determined based on the information entropy matrix corresponding to the first decomposition graph. The specific process is as follows: based on the attention mechanism, the importance coefficient of the information entropy matrix is ​​calculated; based on the importance coefficient of the information entropy matrix, the importance coefficient corresponding to the first heterogeneous graph is calculated.

[0109] In this embodiment of the application, the second information entropy and the second weight corresponding to the second heterogeneous graph are determined based on the information entropy weighting method. The specific process is as follows:

[0110] Based on the information entropy weighting method, the interaction information entropy between users and points of interest in the second decomposition graph is determined; based on the information entropy weighting method, the connection information entropy corresponding to the second decomposition graph is determined; based on the interaction information entropy and the connection information entropy, the information entropy matrix corresponding to the second decomposition graph is constructed; based on the information entropy matrix corresponding to the second decomposition graph, the second information entropy and the second weight corresponding to the second heterogeneous graph are determined.

[0111] In this embodiment of the application, the second weight corresponding to the second heterogeneous graph is determined based on the information entropy matrix corresponding to the second decomposition graph. The specific process is as follows:

[0112] Based on the attention mechanism, the importance coefficient of the information entropy matrix is ​​calculated; based on the importance coefficient of the information entropy matrix, the importance coefficient corresponding to the second heterogeneous graph is calculated.

[0113] It should be noted that the method for determining the first information entropy and the first weight is the same as the method for determining the second information entropy and the second weight. For ease of explanation, only one example will be used here. Specifically,

[0114] First, calculate the information entropy of user-point of interest interaction and the connection information entropy between users and between points of interest in the decomposed graph under different association patterns. The calculation formulas are as follows:

[0115] ;

[0116] ;

[0117] in, For nodes The number of neighboring nodes, for The degree, For logarithmic operations, For association mode The node interaction matrix below, where each element represents the number of interactions. If the number of interactions is 0, then the entropy value is 0.

[0118] exist The information entropy matrix is ​​constructed by obtaining the entropy values ​​of all nodes under each association pattern. , This is the embedding dimension. Then, a non-linear transformation is performed on the information entropy matrix using a single MLP layer, as shown in the formula:

[0119] ;

[0120] It is a learnable weight matrix. This is a bias term. Finally, an attention mechanism is introduced to adjust the importance of nodes for different association patterns. The calculation formula is:

[0121] ;

[0122] in It is the hidden layer dimension. It is a learnable weight matrix. It is a bias vector. It is a parameter vector, and the weights are normalized using the softmax function to obtain the importance coefficient of each association pattern. The calculation method is as follows:

[0123] ;

[0124] Where m is the number of association patterns.

[0125] We can obtain user embedding vectors and interest point embedding vectors by weighting the embedding representations of users and interest points using importance coefficients. The calculation method is as follows:

[0126] , ;

[0127] in, For the first User embedding matrix, For the first An interest point embedding matrix.

[0128] This application provides an interest point recommendation device based on association patterns and heterogeneous dual graphs, such as... Figure 2 As shown, it includes:

[0129] The acquisition module 201 is used to acquire a set of user information, a set of point of interest information, a set of point of interest category information, and a set of user check-in time information.

[0130] Construction module 202 is used to construct at least one association pattern based on the user information set, the interest point information set, the interest point category information set, and the user check-in time information set. The association pattern is used to represent the association relationship between user information, interest point information, interest point category information, and user check-in time information. Based on the user information set, the interest point information set, the interest point category information set, and the user check-in time information set, a first heterogeneous graph of interest point information based on semantic relationship and a second heterogeneous graph of user information based on time association are constructed respectively.

[0131] The data processing module 203 is used to determine, based on the association pattern and vector embedding, a first decomposition graph of the first heterogeneous graph and a second decomposition graph of the second heterogeneous graph, respectively; to determine the first information entropy and the first weight corresponding to the first heterogeneous graph based on the information entropy weighting method; to determine the second information entropy and the second weight corresponding to the second heterogeneous graph based on the information entropy weighting method; and to determine the recommendation score of each interest point based on the first information entropy, the first weight, the second information entropy, and the second weight.

[0132] This application provides a storage medium for storing computer-executable instructions, characterized in that, when executed, the computer-executable instructions implement the steps of the interest point recommendation method based on association patterns and heterogeneous dual graphs as described in any one of the embodiments.

[0133] It should be noted that the embodiments concerning storage media in this application and the embodiments concerning the interest point recommendation method based on association patterns and heterogeneous dual graphs in this application are based on the same inventive concept. Therefore, the specific implementation of this embodiment can refer to the aforementioned implementation of the corresponding interest point recommendation method based on association patterns and heterogeneous dual graphs, and the repeated parts will not be described again.

[0134] The foregoing has described specific embodiments of this application. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims may be performed in a different order than that shown in the embodiments and may still achieve the desired results. Furthermore, the processes depicted in the drawings do not necessarily require the specific or sequential order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0135] In the 1930s, improvements to a technology could be clearly distinguished as either hardware improvements (e.g., improvements to the circuit structure of diodes, transistors, switches, etc.) or software improvements (improvements to the methodology). However, with technological advancements, many improvements to the methodology today can be considered direct improvements to the hardware circuit structure. Designers almost always obtain the corresponding hardware circuit structure by programming the improved methodology into the hardware circuit. Therefore, it cannot be said that an improvement to the methodology cannot be implemented using hardware physical modules. For example, a Programmable Logic Device (PLD) (such as a Field Programmable Gate Array (FPGA)) is such an integrated circuit whose logic function is determined by the user programming the device. Designers can program and "integrate" a digital system onto a PLD themselves, without needing chip manufacturers to design and manufacture dedicated integrated circuit chips. Furthermore, nowadays, instead of manually manufacturing integrated circuit chips, this programming is mostly implemented using "logic compiler" software. Similar to the software compiler used in program development, the original code before compilation must also be written in a specific programming language, called a Hardware Description Language (HDL). There are many HDLs, such as ABEL (Advanced Boolean Expression Language), AHDL (Altera Hardware Description Language), Confluence, CUPL (Cornell University Programming Language), HDCal, JHDL (Java Hardware Description Language), Lava, Lola, MyHDL, PALASM, and RHDL (Ruby Hardware Description Language). Currently, the most commonly used are VHDL (Very-High-Speed ​​Integrated Circuit Hardware Description Language) and Verilog. Those skilled in the art should also understand that by simply performing some logic programming on the method flow using one of these hardware description languages ​​and programming it into an integrated circuit, the hardware circuit implementing the logical method flow can be easily obtained.

[0136] The controller can be implemented in any suitable manner. For example, it can take the form of a microprocessor or processor and a computer-readable medium storing computer-readable program code (e.g., software or firmware) executable by the (micro)processor, logic gates, switches, application-specific integrated circuits (ASICs), programmable logic controllers, and embedded microcontrollers. Examples of controllers include, but are not limited to, the following microcontrollers: ARC 625D, Atmel AT91SAM, Microchip PIC18F26K20, and Silicon Labs C8051F320. A memory controller can also be implemented as part of the control logic of the memory. Those skilled in the art will also recognize that, in addition to implementing the controller in purely computer-readable program code form, the same functionality can be achieved by logically programming the method steps to make the controller take the form of logic gates, switches, application-specific integrated circuits, programmable logic controllers, and embedded microcontrollers. Therefore, such a controller can be considered a hardware component, and the means included therein for implementing various functions can also be considered as structures within the hardware component. Alternatively, the means for implementing various functions can be considered as both software modules implementing the method and structures within the hardware component.

[0137] The systems, devices, modules, or units described in the above embodiments can be implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer. Specifically, a computer can be, for example, a personal computer, laptop computer, cellular phone, camera phone, smartphone, personal digital assistant, media player, navigation device, email device, game console, tablet computer, wearable device, or any combination of these devices.

[0138] For ease of description, the above apparatus is described by dividing it into various functional units. Of course, in implementing the embodiments of this application, the functions of each unit can be implemented in one or more software and / or hardware.

[0139] Those skilled in the art will understand that one or more embodiments of this application can be provided as a method, system, or computer program product. Therefore, one or more embodiments of this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can 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, etc.) containing computer-usable program code.

[0140] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can 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, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0141] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0142] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0143] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0144] Memory may include non-persistent storage in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0145] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0146] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0147] One or more embodiments of this application can be described in the general context of computer-executable instructions, such as program modules, that are executed by a computer. Generally, program modules include routines, programs, objects, components, data structures, etc., that perform a specific task or implement a specific abstract data type. One or more embodiments of this application can also be practiced in distributed computing environments where tasks are performed by remote processing devices connected via a communication network. In a distributed computing environment, program modules can reside in local and remote computer storage media, including storage devices.

[0148] The various embodiments in this application are described in a progressive manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the system embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions of the method embodiments.

[0149] The above description is merely an embodiment of this document and is not intended to limit the scope of this document. Various modifications and variations can be made to this document by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this document should be included within the scope of the claims of this document.

Claims

1. A method for recommending points of interest based on association patterns and heterogeneous dual graphs, characterized in that, include: Obtain a set of user information, a set of points of interest information, a set of points of interest category information, and a set of user check-in time information; Based on the user information set, the point of interest information set, the point of interest category information set, and the user check-in time information set, at least one association pattern is constructed, which is used to represent the association relationship between user information, point of interest information, point of interest category information, and user check-in time information. Based on the association pattern, the user information set, the interest point information set, the interest point category information set, and the user check-in time information set, a first heterogeneous graph of interest point information based on semantic relationships and a second heterogeneous graph of user information based on time association are constructed respectively. Based on the association pattern, and using vector embedding, the first decomposition graph of the first heterogeneous graph and the second decomposition graph of the second heterogeneous graph are determined respectively. Based on the information entropy weighting method, the first information entropy and the first weight corresponding to the first heterogeneous graph are determined according to the first decomposition graph. Based on the information entropy weighting method, the second information entropy and the second weight corresponding to the second heterogeneous graph are determined according to the second decomposition graph. The recommendation score for each point of interest is determined based on the first information entropy, the first weight, the second information entropy, and the second weight.

2. The method according to claim 1, characterized in that, After obtaining the exploded graph, the method further includes: Add negative samples to the first or second decomposition plot.

3. The method according to claim 2, characterized in that, Adding negative samples includes: Set the total number of training rounds, the initial proportion of negative samples, and the maximum proportion of negative samples; For each training round, the current negative sample ratio is adjusted based on the linear law and the initial negative sample ratio until the current negative sample ratio equals the highest negative sample ratio.

4. The method according to claim 3, characterized in that, Adjusting the current negative sample ratio based on the linear law and the initial negative sample ratio includes: The current proportion of negative samples is calculated using the following formula: ； in, This represents the current proportion of negative samples. For the total number of training rounds, Indicates the current training round. For the initial ratio and It represents the highest proportion.

5. The method according to claim 1, characterized in that, Based on the information entropy weighting method, the first information entropy and first weight corresponding to the first heterogeneous graph are determined according to the first decomposition graph, including: Based on the information entropy weighting method, the information entropy of user-interest point interaction in the first decomposition graph is determined. Based on the information entropy weighting method, the connection information entropy corresponding to the first decomposition graph is determined; Based on the interaction information entropy and the connection information entropy, construct the information entropy matrix corresponding to the first decomposition graph; Based on the information entropy matrix corresponding to the first decomposition graph, determine the first information entropy and the first weight corresponding to the first heterogeneous graph.

6. The method according to claim 5, characterized in that, Based on the information entropy matrix corresponding to the first decomposition graph, the first weight corresponding to the first heterogeneous graph is determined, including: The importance coefficient of the information entropy matrix is ​​calculated based on the attention mechanism; The importance coefficient corresponding to the first heterogeneous graph is calculated based on the importance coefficient of the information entropy matrix.

7. The method according to claim 1, characterized in that, Based on the information entropy weighting method, the second information entropy and second weight corresponding to the second heterogeneous graph are determined according to the second decomposition graph, including: Based on the information entropy weighting method, the information entropy of user-interest point interaction in the second decomposition graph is determined; Based on the information entropy weighting method, the connection information entropy corresponding to the second decomposition graph is determined; Based on the interaction information entropy and the connection information entropy, construct the information entropy matrix corresponding to the second decomposition graph; Based on the information entropy matrix corresponding to the second decomposition graph, determine the second information entropy and the second weight corresponding to the second heterogeneous graph.

8. The interest point recommendation method based on association patterns and heterogeneous dual graphs according to claim 7, characterized in that, Based on the information entropy matrix corresponding to the second decomposition graph, the second weights corresponding to the second heterogeneous graph are determined, including: The importance coefficient of the information entropy matrix is ​​calculated based on the attention mechanism; The importance coefficient of the second heterogeneous graph is calculated based on the importance coefficient of the information entropy matrix.

9. An interest point recommendation device based on association patterns and heterogeneous dual graphs, characterized in that, include: The acquisition module is used to acquire user information sets, interest point information sets, interest point category information sets, and user check-in time information sets. The construction module is used to construct at least one association pattern based on the user information set, the interest point information set, the interest point category information set, and the user check-in time information set. The association pattern is used to represent the association relationship between user information, interest point information, interest point category information, and user check-in time information. Based on the user information set, the interest point information set, the interest point category information set, and the user check-in time information set, the module also constructs a first heterogeneous graph of interest point information based on semantic relationships and a second heterogeneous graph of user information based on time associations. The data processing module is used to determine, based on the association pattern and vector embedding, the first decomposition graph of the first heterogeneous graph and the second decomposition graph of the second heterogeneous graph, respectively. Based on the information entropy weighting method, the first information entropy and the first weight corresponding to the first heterogeneous graph are determined; Based on the information entropy weighting method, the second information entropy and the second weight corresponding to the second heterogeneous graph are determined. The recommendation score for each point of interest is determined based on the first information entropy, the first weight, the second information entropy, and the second weight.

10. A storage medium for storing computer-executable instructions, characterized in that, When executed, the computer-executable instructions implement the steps of the interest point recommendation method based on association patterns and heterogeneous dual graphs as described in any one of claims 1-8.

Images