Diffusion map recommendation method and system based on multi-hop mechanism

By constructing a knowledge graph and POI-GCN feature extraction network, combining water wave diffusion and multi-hop mechanisms, the limitations of existing recommendation algorithms in processing complex semantic information are solved, and more accurate user interest understanding and personalized recommendation are achieved.

CN118981545AActive Publication Date: 2024-11-19DALIAN OCEAN UNIV
View PDF 5 Cites 0 Cited by

Patent Information

Application Number
CN202411191785.5
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-28
Publication Date
2024-11-19
Estimated Expiration
2044-08-28

AI Technical Summary

Technical Problem

Existing recommendation algorithms have limitations in processing complex semantic information and contextual information, and cannot fully understand the interests and needs of users, and there are problems such as difficulty in advanced modeling and insufficient user feature modeling.

Method used

Using the diffusion map recommendation method based on the multi-hop mechanism, the knowledge map, POI-GCN feature extraction network and the Diffussion-POI recommendation model are constructed, and the deep semantic information in the knowledge map is captured, and the recommendation sequence is generated through the water wave diffusion and multi-hop mechanism. .

Benefits of technology

The model's expression ability and generalization ability can better capture the characteristics of different types of interest points, enrich user portraits, discover potential relationships, and achieve personalized and diversified recommendations.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118981545A_ABST
    Figure CN118981545A_ABST
Patent Text Reader

Abstract

The invention provides a diffusion graph recommendation method and system based on a multi-hop mechanism, and belongs to the technical field of deep learning, knowledge graphs and recommendation systems. The method is composed of knowledge graph construction, a feature extraction network and a multi-hop mechanism diffusion model. Different types of POI features are respectively input into different feature extraction hidden layers, and different operations such as weight distribution, embedded mapping and nonlinear transformation are performed on the POI features, so that feature differences of different types of interest points are better captured, and the expression ability and generalization ability of the model are improved; through information transmission of the graph, user portraits are enriched, potential association is found, and personalized recommendation is carried out; the water wave diffusion algorithm can enable the model to have the capability of searching for more entities so as to improve the generalization performance of the model; the multi-hop model can gradually expand and enrich interest point sets of the user, more potential interest points related to historical behaviors of the user are mined, relation information in the user and the knowledge graph is effectively utilized, and personalized and diversified recommendation is achieved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention belongs to the technical field of deep learning, knowledge graphs and recommendation systems, and specifically relates to a diffusion graph recommendation method and system based on a multi-hop mechanism. Background Art

[0002] With the rapid development of information technology and the widespread popularity of Internet applications, users are faced with a vast amount of information and choices in their daily lives. In this era of information explosion, in order to accurately obtain information that users are interested in, points of interest are often used to describe user information, preferences and other personal information. As an important branch of the recommendation system, point of interest recommendation aims to recommend content that users may be interested in based on information such as users' interests, preferences and historical behaviors, thereby helping users to obtain useful information more efficiently. Recommendation algorithms have emerged. With the rise of knowledge graph technology, point of interest recommendation algorithms based on knowledge graphs have gradually attracted attention. Knowledge graphs are a semantic network that can describe concepts, entities and the relationships between them in the real world, providing rich semantic information and contextual information for recommendation systems. By combining knowledge graphs with recommendation algorithms, users' interests and needs can be understood more accurately, and the accuracy and diversity of recommendations can be improved.

[0003] In recent years, personalized recommendation systems and knowledge graphs have made significant progress in academia and industry. Traditional recommendation algorithms mainly make recommendations based on users' historical behavior data, such as collaborative filtering and content filtering. The existing technology introduces rating preference models and time weight factors, reconstructs the user-item matrix, and optimizes the traditional collaborative filtering algorithm, but there is still the problem of incomplete user preference mining. To address this problem, Li Xiangkun et al. proposed a collaborative filtering method (OCRIF) that integrates overlapping community regularization and implicit feedback. This method takes into account the community structure of users in the community network and the implicit feedback of user rating information and social information, further improving the performance of collaborative filtering algorithms. However, these methods have limitations in processing complex semantic information and contextual information, and cannot fully understand users' interests and needs. They also have common problems such as difficulty in high-order modeling and insufficient user feature modeling. Summary of the invention

[0004] To solve the above problems, the present invention proposes a diffusion graph recommendation method based on a multi-hop mechanism, which mines high-order semantic information in the knowledge graph in an end-to-end manner. It covers three parts: knowledge graph construction, feature extraction network construction, and multi-hop diffusion model construction, specifically including:

[0005] Construct a knowledge graph that contains basic information about various points of interest and incorporates semantic information and social relationships;

[0006] A POI-GCN feature extraction network is constructed based on a graph convolutional neural network to capture deep semantic information in the knowledge graph;

[0007] Based on the features extracted by the POI-GCN feature extraction network, water wave diffusion and multi-hop mechanism are introduced to construct a Diffussion-POI recommendation model to generate a recommendation sequence.

[0008] Furthermore, the method for constructing the knowledge graph is:

[0009] The crawled Yelp2018 dataset, Gowalla dataset, public business address data, and open platform merchant data obtained using the MeiTuan-POI interface are used to form a knowledge graph dataset. The knowledge graph dataset consists of structured point of interest data and unstructured and semi-structured data, providing high-quality data for subsequent model training;

[0010] Performing data preprocessing on the data in the knowledge graph dataset;

[0011] The triples <entity, relationship, attribute> that make up the knowledge graph are constructed based on the preprocessed text results.

[0012] Furthermore, the POI-GCN feature extraction network includes a word embedding layer, a network layer, a GCN layer and a hidden layer;

[0013] For each point of interest, extract features including the name, location, category, and score of the point of interest from its triplet, and then map these features into vectors of fixed length and input them into the word embedding layer to obtain an embedding vector of the entity as the input of the network layer;

[0014] The attention mechanism is added to the network layer to distribute weights. The calculation formula is as follows:

[0015]

[0016] Among them, a ij represents the attention weight; softplus is the activation function; e ij Represents the similarity between user i and item j. The formula is as follows:

[0017] e ij =ELU(a T |Wh r ||Wh i |) (2)

[0018]

[0019] Among them, ELU is the activation function; W is the linear transformation matrix, the shape is W∈RF×F′ , F is the dimension of the input feature, F' is the dimension of the output feature; Wh r Represents the element in the hth row and rth column of the W matrix; Wh i represents the element in the hth row and ith column of the W matrix; a represents the parameter vector; f(x) represents the ELU activation function, which adjusts the similarity value e through nonlinear transformation ij , to better calculate the attention weight; exp represents the natural exponential function; γ is the scaling factor;

[0020] After weight assignment, the weighted entity embedding vector is output, and then message aggregation is performed to obtain aggregate features;

[0021] The GCN layer performs feature fusion on the obtained aggregated features and outputs an entity feature vector of the fused features;

[0022] The hidden layer uses the ReLU activation function to perform nonlinear transformation processing on the entity feature vector of the fused feature to obtain a feature vector that integrates all entity information.

[0023] Furthermore, the water wave diffusion and multi-hop mechanism are introduced to construct the Diffussion-POI recommendation model, and the process of generating the recommendation sequence is as follows:

[0024] S10: Extracting knowledge graph features based on the feature vector extracted by the POI-GCN feature extraction network

[0025] S101: Perform entity linking on entities in the knowledge graph to obtain matched entities;

[0026] S102: extracting subgraphs from the knowledge graph after entity linking;

[0027] S103: Introducing contextual text features into the entities in the subgraph:

[0028] context(n)={n m |(n,r,n m )∈G or(n m ,r,n)∈G}

[0029] Where context(n) is a function used to describe the contextual features of entity n; G represents the graph; r represents the relationship between nodes; n m represents the mth entity; (n,r,n m ) represents a triple, entity n, relation r and another entity n m ;

[0030] S20: Transmitting graph information

[0031] S201: setting the initial value of all entity nodes after the contextual text feature is introduced to 0, and setting the initial value of the seed node to 1;

[0032] S202: Iteratively update the value of each entity node. In each iteration, for any entity node v, calculate the contribution value d(x, v) of its neighbor node x to it, and then update the value of v to the sum of d(x, v), so as to continuously spread;

[0033] S203: Repeat the above iterative update process until the values ​​of all entity nodes no longer change or the preset maximum number of iterations is reached, and output feature representations of all entities;

[0034] S30: Perform multi-hop recommendation

[0035] S301: Perform one-hop reasoning: input the feature representation of the entity, the feature vector integrating all entity information and the user's historical behavior and convert them into an embedding vector; randomly select a node in the subgraph as the starting node, and calculate the transition probability of the relationship from the starting node to its direct neighbor node; calculate the similarity between the user embedding vector and the entity embedding vector in the subgraph; then use the transition probability to weight the similarity; then diffuse through iteration to simulate the random walk of the user in the subgraph; after multiple rounds of iterations, find the entity set that is highly correlated with the user's historical behavior after one hop;

[0036] S302: Perform two-hop reasoning: randomly select a new starting point from the obtained entity set; then calculate the two-hop transfer probability, combine the user embedding vector and the two-hop transfer probability, and find the entity set that is more deeply associated with the user's historical behavior after two hops;

[0037] S303: Repeat the above process for multiple iterations to obtain multiple entity sets with different degrees of association with the user's historical behavior, perform weighted summation on the multiple entity sets to obtain a weighted entity set, generate a recommendation sequence for the entities ranked first in weight value, and implement personalized recommendation.

[0038] Furthermore, the data preprocessing includes:

[0039] Perform data cleaning to remove irrelevant information;

[0040] Perform Chinese word segmentation on the cleaned data;

[0041] Perform entity recognition on the data after Chinese word segmentation, and the recognized place names are the points of interest, and obtain the entity data of the points of interest: for structured data, the place name entities are extracted based on the Bi-LSTM-CRF algorithm; for semi-structured data, a rule-based method is used to classify and label the specific address information;

[0042] Based on spatial geographic information, the identified place name entities are disambiguated, extracted and aligned to obtain the point of interest corpus;

[0043] Based on dependency syntactic analysis, the relationship of the interest point corpus is extracted, and the relationship of the interest points is divided to obtain the final corpus text data.

[0044] Furthermore, an AC automaton based on a double-array dictionary tree is used for Chinese word segmentation.

[0045] Furthermore, the Bert-BiLSTM-CRF neural network algorithm is used for entity extraction and alignment, and the KMP algorithm is used to improve the efficiency of entity alignment.

[0046] Furthermore, the interest point relationship is divided into: subject-predicate-object, verb-complement structure, adverbial-verb structure, adverbial-verb-complement structure, attributive postposition relationship, preposition-object relationship, and object preposition structure.

[0047] The present invention also provides a diffusion graph recommendation system based on a multi-hop mechanism, including a knowledge graph construction module, a POI-GCN feature extraction network and a Diffussion-POI recommendation model;

[0048] The knowledge graph construction module is used to construct a knowledge graph containing basic information of various points of interest and integrating semantic information and social relationships;

[0049] The POI-GCN feature extraction network is used to capture deep semantic information in the knowledge graph;

[0050] The Diffussion-POI recommendation model is used to generate a recommendation sequence by performing personalized recommendation based on the features extracted by the POI-GCN feature extraction network and combining water wave diffusion with a multi-hop mechanism.

[0051] Beneficial effects of the present invention: The present invention provides a diffusion graph recommendation method and system based on a multi-hop mechanism, which is mainly composed of knowledge graph construction, feature extraction network construction and multi-hop mechanism diffusion model construction; different types of POI features are respectively input into different feature extraction hidden layers, and different weight allocations, embedding mappings and nonlinear transformations are performed on them, which can better capture the feature differences of different types of interest points, thereby improving the expression and generalization capabilities of the model; through the information transmission of the graph, user portraits can be enriched, potential associations can be discovered and personalized recommendations can be made; the water wave diffusion algorithm can enable the model to have the ability to find more entities to improve its generalization performance; the multi-hop model can gradually expand and enrich the user's interest point set, explore more potential interest points related to the user's historical behavior, and effectively utilize the relationship information between users and knowledge graphs to achieve personalized and diversified recommendations. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 A schematic diagram of a flow chart of a diffusion graph recommendation method based on a multi-hop mechanism provided by an embodiment of the present invention;

[0053] Figure 2 The overall architecture diagram of the diffusion graph recommendation model based on the multi-hop mechanism provided by the embodiment of the present invention;

[0054] Figure 3 A schematic diagram of the process of constructing a knowledge graph provided by an embodiment of the present invention;

[0055] Figure 4 The overall architecture diagram of the feature extraction network model provided by the embodiment of the present invention;

[0056] Figure 5 A calculation structure diagram of Wighted Sum provided in an embodiment of the present invention;

[0057] Figure 6 A schematic diagram of a multi-hop model provided by an embodiment of the present invention; DETAILED DESCRIPTION

[0058] The specific implementation of the present invention is further described in detail below in conjunction with the accompanying drawings and examples. The following examples are used to illustrate the present invention, but the protection scope of the present invention is not limited to the following examples.

[0059] like Figure 1 and Figure 2 As shown, the embodiment of the present invention provides a diffusion graph recommendation method based on a multi-hop mechanism, including three parts: knowledge graph construction, feature submission network construction, and multi-hop mechanism diffusion model construction. The specific steps are as follows:

[0060] Step 1: Build a knowledge graph, see Figure 3 ;

[0061] Step 1.1: The crawled Yelp2018 dataset, Gowalla dataset, AutoNavi POI and other public commercial address data, as well as the open platform merchant data obtained using the MeiTuan-POI interface, are used to form a knowledge graph dataset. The knowledge graph dataset consists of structured point of interest data and unstructured and semi-structured data, providing high-quality data for subsequent model training.

[0062] Step 1.2: Preprocess the data in the knowledge graph dataset

[0063] Step 1.2.1: Use regular expressions, string matching and other methods to clean the data and remove irrelevant information such as noise data, non-text characters, stop words, etc.

[0064] Step 1.2.2: Use AC automaton based on double-array dictionary tree to perform Chinese word segmentation on the cleaned data to achieve efficient multi-mode matching, which can reduce the number of comparisons as much as possible and improve matching efficiency;

[0065] Step 1.2.3: Perform entity recognition on the data after Chinese word segmentation. The recognized place names are the points of interest, and the entity data of the points of interest are obtained: for structured data, the place name entities are extracted based on the Bi-LSTM-CRF algorithm; for semi-structured data, the specific address information is classified and labeled using a rule-based method;

[0066] Step 1.2.4: Based on spatial geographic information, entity disambiguation, entity extraction and entity alignment are performed on the identified place name entities to obtain the point of interest corpus; the Bert-BiLSTM-CRF neural network algorithm is used for entity extraction and alignment, and the KMP algorithm is used to improve the efficiency of entity alignment;

[0067] Step 1.2.5: Extract relations from the interest point corpus based on dependency syntactic analysis, and divide the interest point relations into seven types: subject-predicate-object, verb-complement structure, adverbial-verb structure, adverbial-verb-complement structure, post-attributive relation, preposition-object relation, and object-preposition structure, so as to facilitate relation extraction and obtain the final corpus text data.

[0068] Step 1.3: Construct the triples <entity, relationship, attribute> that make up the knowledge graph based on the preprocessed text results, store the knowledge graph in the repository neo4j and visualize it. The knowledge graph not only contains the basic information of various points of interest, but also incorporates rich semantic information and social relationships, providing users with more comprehensive and in-depth knowledge services.

[0069] Step 2: Based on the graph convolutional neural network (GCN), a feature extraction network based on POI-GCN is constructed to capture the deep semantic information in the knowledge graph and integrate it into the recommendation system to enhance the recommendation effect. The triples in the knowledge graph are used as semantic networks, with nodes representing entities and edges representing various semantic relationships between entities or between entities and attributes.

[0070] Figure 4 The overall architecture diagram of the feature extraction network includes a word embedding layer, a network layer, a GCN layer, and a hidden layer;

[0071] For each point of interest, extract features including the name, location, category, and score of the point of interest from its triplet, and then map these features into vectors of fixed length and input them into the word embedding layer to obtain an embedding vector of the entity as the input of the network layer;

[0072] The attention mechanism is added to the network layer to distribute weights. The calculation formula is as follows:

[0073]

[0074] Among them, a ij represents the attention weight; softplus is the activation function; e ij Represents the similarity between user i and item j. The formula is as follows:

[0075] e ij =ELU(a T |Wh r ||Wh i |) (2)

[0076]

[0077] Among them, ELU is the activation function; W is the linear transformation matrix, the shape is W∈R F×F′ , F is the dimension of the input feature, F' is the dimension of the output feature; Wh r Represents the element in the hth row and rth column of the W matrix; Wh i represents the element in the hth row and ith column of the W matrix; a represents the parameter vector; f(x) represents the ELU activation function, which adjusts the similarity value e through nonlinear transformation ij , to better calculate the attention weight; exp represents the natural exponential function; γ is the scaling factor;

[0078] After weight assignment, the weighted entity embedding vector is output, and then message aggregation is performed to obtain the aggregated features;

[0079] The GCN layer performs feature fusion on the obtained aggregated features and outputs an entity feature vector of the fused features;

[0080] The hidden layer uses the ReLU activation function to perform nonlinear transformation processing on the entity feature vector of the fused feature, solves the problems of gradient disappearance and neuron death, and captures the feature vector that integrates all entity information.

[0081] Step 3: Based on the features extracted by the POI-GCN feature extraction network, water wave diffusion and multi-hop mechanisms are introduced in an end-to-end manner to build a Diffussion-POI recommendation model for recommendation.

[0082] Step 3.1: Extract knowledge graph features based on the feature vector extracted by the POI-GCN feature extraction network

[0083] Step 3.1.1: Perform entity linking on the entities in the knowledge graph to obtain matched entities;

[0084] Step 3.1.2: Extract subgraphs from the knowledge graph after entity linking;

[0085] Step 3.1.3: Introduce contextual text features to the entities in the subgraph to more accurately characterize the features of the entities. The formula is as follows:

[0086] context(n)={n m |(n,r,n m )∈G or(n m ,r,n)∈G}

[0087] Where context(n) is a function used to describe the context features of entity n; G represents the graph; r represents the relationship between nodes; m represents the mth entity; (n, r, n m ) represents a triple, entity n, relation r and another entity n m .

[0088] Step 3.2: Transfer graph information

[0089] Step 3.2.1: Set the initial value of all entity nodes after the contextual text features are introduced to 0, and set the initial value of the seed node to 1;

[0090] Step 3.2.2: Iterate and update the value of each entity node. In each iteration, for any entity node v, calculate the contribution value d(x,v) of its neighbor node x to it, and then update the value of v to the sum of d(x,v), so as to continuously spread;

[0091] Step 3.2.3: Repeat the above iterative update process until the values ​​of all entity nodes no longer change or the preset maximum number of iterations is reached; output the feature representation of all entities.

[0092] Step 3.3: Perform multi-hop recommendation, such as Figure 6 As shown, the specific process is as follows:

[0093] Step 3.3.1: Perform one-hop reasoning. The specific process is as follows: input the feature representation of the entity, integrate the feature vector of all entity information and the user's historical behavior and convert it into an embedding vector; randomly select a node in the subgraph as the starting node, and calculate the transition probability of the relationship from the starting node to its direct neighbor node; calculate the similarity between the user embedding vector and the entity embedding vector in the subgraph; then use the transition probability to weight the similarity; then diffuse through iteration to simulate the random walk of the user in the subgraph; after multiple rounds of iterations, find the entity set that is highly correlated with the user's historical behavior after one hop;

[0094] Step 3.3.2: Perform two-hop reasoning. The specific process is as follows: after completing the first hop, randomly select a new starting point from the obtained interest point set; then calculate the two-hop transfer probability, combine the user embedding vector and the two-hop transfer probability, and find the entity set that is more deeply associated with the user's historical behavior after two hops;

[0095] Step 3.3.3: Repeat the above process for multiple iterations to obtain multiple entity sets with different degrees of association with the user's historical behavior. Use Weight Sum to perform weighted summation on the multiple entity sets. The calculation process is shown in Figure 5 , obtain the weighted entity set, generate a recommendation sequence for the entities with the highest weight values, and realize personalized recommendation.

[0096] Based on the above method, some hyperparameters are tested to verify the robustness of the model.

[0097] 1) Set the linear transformation layer to linear and sigmid, the masking ratio to 0.2, the learning rate to 0.01, the hidden layer dimension to 64, the number of propagation layers to 2, and the other model settings to be constant. Adjust the adjacency matrix propagation layer comparison dimension settings. The experimental results are shown in Table 1:

[0098] Table 1 Adjacency matrix layer dimension settings

[0099]

[0100] Experimental results show that the dimension setting of the model is a key parameter, and 512-dimensional feature representation has the best effect.

[0101] 2) Set the linear transformation layer to linear and sigmid, the masking ratio to 0.2, the learning rate to 0.01, the hidden layer dimension to 64, and the other model settings to remain constant. Adjust the adjacency matrix propagation layer comparison dimension settings. The experimental results are shown in Table 2:

[0102] Table 2 Propagation layer depth settings

[0103]

[0104] In this example, based on the characteristics of the specific task and data set, the number of propagation layers is adjusted to 6 to achieve the best performance.

[0105] The recommended model of the present invention is named MultiHop-GDN model, and an ablation experiment is performed. The experimental results are shown in Table 3.

[0106] Table 3. Comparison of ablation experiment performance

[0107]

[0108] A performance comparison experiment was conducted, considering the accuracy, recall and F1 value on three datasets: Gowalla, Yelp2018, and MeiTuan-POI as evaluation indicators.

[0109] To ensure the quality of the Gowalla dataset, this example uses a 10-core setting, that is, users and items are retained through at least ten interactions. The Yelp2018 dataset is taken from the 2018 version of the Yelp Challenge. Among them, local businesses such as restaurants and bars are regarded as items, and the same 10-core setting is used to ensure data quality. For the MeiTuan-POI dataset, the open platform merchant categories of life services, transportation hubs, and medical health are selected from the collection. Similarly, a 10-core setting is used to ensure that each user and item has at least 10 interactions. This example is compared with the four models of Pinsage, NGCF, KM-CF, and BC-CF. The experimental results are shown in Table 4.

[0110] Table 4 Overall comparison of performance (RQ1)

[0111]

[0112] The experimental results on the public datasets verify that the MultiHop-GDN model of the present invention has better recommendation effects than other baseline methods, showing the powerful ability of the model in capturing user and project characteristics. The present invention provides a useful reference for further integrating knowledge graphs and deepening recommendation mechanisms in the field of recommendation systems in the future. Future research can explore the application of multi-hop mechanisms in a wider range of recommendation system types, as well as more efficient methods for integrating knowledge graphs with recommendation systems, to further improve the overall performance of recommendation systems and user satisfaction.

[0113] Finally, it should be noted that the above embodiments are intended to illustrate the technical solutions of the present invention and do not constitute any form of limitation on the present invention. Those skilled in the art should fully understand that it is completely feasible to modify the technical solutions described in the above embodiments or to perform equivalent replacements on any part or all of the technical features therein. These modifications or replacements, as long as they do not deviate from the scope of protection determined by the claims of the present invention, should be regarded as reasonable extensions of the present invention.

Claims

1. A diffusion graph recommendation method based on multi-hop mechanism, It is characterized in that include: Construct a knowledge graph that contains basic information about various points of interest and incorporates semantic information and social relationships; A POI-GCN feature extraction network is constructed based on a graph convolutional neural network to capture deep semantic information in the knowledge graph; Based on the features extracted by the POI-GCN feature extraction network, water wave diffusion and multi-hop mechanism are introduced to construct a Diffussion-POI recommendation model to generate a recommendation sequence.

2. According to claim 1, a diffusion graph recommendation method based on a multi-hop mechanism, It is characterized in that The method for constructing the knowledge graph is as follows: The crawled Yelp2018 dataset, Gowalla dataset, public business address data, and open platform merchant data obtained using the MeiTuan-POI interface are used to form a knowledge graph dataset, which consists of structured point of interest data and unstructured and semi-structured data; Performing data preprocessing on the data in the knowledge graph dataset; The triples <entity, relationship, attribute> that make up the knowledge graph are constructed based on the preprocessed text results.

3. According to claim 2, a diffusion graph recommendation method based on a multi-hop mechanism, It is characterized in that The POI-GCN feature extraction network includes a word embedding layer, a network layer, a GCN layer and a hidden layer; For each point of interest, extract features including the name, location, category, and score of the point of interest from its triplet, and then map these features into vectors of fixed length and input them into the word embedding layer to obtain an embedding vector of the entity as the input of the network layer; The attention mechanism is added to the network layer to distribute weights. The calculation formula is as follows: Among them, a ij represents the attention weight; softplus is the activation function; e ij Represents the similarity between user i and item j. The formula is as follows: have been ij =TWO(a T |Wh r ||Wh i |) (2) Among them, ELU is the activation function; W is the linear transformation matrix, the shape is W∈R F×F’ , F is the dimension of the input feature, F' is the dimension of the output feature; Wh r Represents the element in the hth row and rth column of the W matrix; Wh i represents the element in the hth row and ith column of the W matrix; a represents the parameter vector; f(x) represents the ELU activation function, which adjusts the similarity value e through nonlinear transformation ij ; exp represents the natural exponential function; γ is the scaling factor; After weight assignment, the weighted entity embedding vector is output, and then message aggregation is performed to obtain aggregate features; The GCN layer performs feature fusion on the obtained aggregated features and outputs an entity feature vector of the fused features; The hidden layer uses the ReLU activation function to perform nonlinear transformation processing on the entity feature vector of the fused feature to obtain a feature vector that integrates all entity information.

4. According to claim 3, a diffusion graph recommendation method based on a multi-hop mechanism, It is characterized in that The water wave diffusion and multi-hop mechanism are introduced to construct the Diffussion-POI recommendation model. The process of generating the recommendation sequence is as follows: S10: Extracting knowledge graph features based on the feature vector extracted by the POI-GCN feature extraction network S101: Perform entity linking on entities in the knowledge graph to obtain matched entities; S102: extracting subgraphs from the knowledge graph after entity linking; S103: Introducing contextual text features into the entities in the subgraph: context (n)={n m |(n,r,n m )∈G or(n m ,r,n)∈G} Where context(n) is a function used to describe the contextual features of entity n; G represents the graph; r represents the relationship between nodes; n m represents the mth entity; (n,r,n m ) represents a triple, entity n, relation r and another entity n m ; S20: Transmitting graph information S201: setting the initial value of all entity nodes after the contextual text feature is introduced to 0, and setting the initial value of the seed node to 1; S202: Iteratively update the value of each entity node. In each iteration, for any entity node v, calculate the contribution value d(x, v) of its neighbor node x to it, and then update the value of v to the sum of d(x, v), so as to continuously spread; S203: Repeat the above iterative update process until the values ​​of all entity nodes no longer change or the preset maximum number of iterations is reached, and output feature representations of all entities; S30: Perform multi-hop recommendation S301: Perform one-hop reasoning: input the feature representation of the entity, the feature vector integrating all entity information and the user's historical behavior and convert them into an embedding vector; randomly select a node in the subgraph as the starting node, and calculate the transition probability of the relationship from the starting node to its direct neighbor node; calculate the similarity between the user embedding vector and the entity embedding vector in the subgraph; then use the transition probability to weight the similarity; then diffuse through iteration to simulate the random walk of the user in the subgraph; after multiple rounds of iterations, find the entity set that is highly correlated with the user's historical behavior after one hop; S302: Perform two-hop reasoning: randomly select a new starting point from the obtained entity set; then calculate the two-hop transfer probability, combine the user embedding vector and the two-hop transfer probability, and find the entity set that is more deeply associated with the user's historical behavior after two hops; S303: Repeat the above process for multiple iterations to obtain multiple entity sets with different degrees of association with the user's historical behavior, perform weighted summation on the multiple entity sets to obtain a weighted entity set, generate a recommendation sequence for the entities ranked first in weight value, and implement personalized recommendation.

5. According to claim 2, a diffusion graph recommendation method based on a multi-hop mechanism, It is characterized in that The data preprocessing includes: Perform data cleaning to remove irrelevant information; Perform Chinese word segmentation on the cleaned data; Perform entity recognition on the data after Chinese word segmentation, and the recognized place names are the points of interest, and obtain the entity data of the points of interest: for structured data, the place name entities are extracted based on the Bi-LSTM-CRF algorithm; for semi-structured data, a rule-based method is used to classify and label the specific address information; Based on spatial geographic information, the identified place name entities are disambiguated, extracted and aligned to obtain the point of interest corpus; Based on dependency syntactic analysis, the relationship of the interest point corpus is extracted, and the relationship of the interest points is divided to obtain the final corpus text data.

6. According to claim 5, a diffusion graph recommendation method based on a multi-hop mechanism, It is characterized in that The AC automaton based on double-array dictionary tree is used for Chinese word segmentation.

7. According to claim 5, a diffusion graph recommendation method based on a multi-hop mechanism, It is characterized in that The Bert-BiLSTM-CRF neural network algorithm is used for entity extraction and alignment, and the KMP algorithm is used to improve the efficiency of entity alignment.

8. According to claim 5, a diffusion graph recommendation method based on a multi-hop mechanism, It is characterized in that The interest point relationship is divided into: subject-predicate-object, verb-complement structure, adverbial-verb structure, adverbial-verb-complement structure, attributive postposition relationship, preposition-object relationship, and object preposition structure.

9. A diffusion graph recommendation system based on a multi-hop mechanism, used to execute the diffusion graph recommendation method according to any one of claims 1 to 8, It is characterized in that Including knowledge graph construction module, POI-GCN feature extraction network and Diffussion-POI recommendation model; The knowledge graph construction module is used to construct a knowledge graph containing basic information of various points of interest and integrating semantic information and social relationships; The POI-GCN feature extraction network is used to capture deep semantic information in the knowledge graph; The Diffussion-POI recommendation model is used to generate a recommendation sequence by performing personalized recommendation based on the features extracted by the POI-GCN feature extraction network and combining water wave diffusion with a multi-hop mechanism.

Citation Information

Patent Citations

  • Intelligent cultural resource recommendation method based on knowledge graph convolutional network

    CN115757897A

  • User interest diffusion recommendation method and system based on knowledge graph

    CN117216281A

  • Service recommendation method and system based on knowledge graph and graph neural network

    CN117688238A

  • Paper recommendation method based on knowledge graph and graph neural network

    CN118364139A

  • Ferroelectric nanoparticle capacitor for non-binary logics

    KR1020240050290A