Tourism route recommendation system based on tourism consumer portrait and psychological model

Abstract

The invention provides a tourism route recommendation system based on tourism consumer portraits and psychological models, the system collects user behavior data through a data acquisition and preprocessing module, and constructs a user and scenic spot interaction information graph, and meanwhile, a scenic spot knowledge graph construction module integrates scenic spot information from top to bottom to form a structured knowledge base, so that the scenic spot information is recommended to the tourism consumer portraits and psychological models. The collaborative knowledge graph construction module combines user behaviors with scenic spot knowledge to provide a basis for the deep learning model and knowledge graph application, the deep learning model application module utilizes a GRU neural network to mine deep features of users and scenic spots, the knowledge graph application module optimizes recommendation results through a TransR model and an attention mechanism, and finally, the recommendation results are recommended to the users and scenic spots. The route planning module plans the travel route according to the information and displays the information to the user through the visualization module, and the system can improve the accuracy and individuation degree of travel route recommendation and improve the user experience.

Description

Technical Field

[0001] The present invention relates to the technical field of tourist route recommendation systems, and in particular to a tourist route recommendation system based on tourist consumer portraits and psychological models. Background Art

[0002] The popularity of the Internet and the widespread use of mobile devices have made it extremely convenient to obtain and share travel information. Tourists are no longer satisfied with traditional group tours or fixed-route tours, but are pursuing a more personalized travel experience. This demand has promoted the rapid development of travel recommendation systems. Initially, route recommendation systems were mainly based on geographic location information and simple user preferences, and recommended travel routes to users through preset algorithms. With the continuous advancement of big data, cloud computing, and artificial intelligence technologies, route recommendation systems have entered a new stage of development. These systems have begun to use massive amounts of tourism data, including user behavior data, attraction information, traffic data, etc., to provide users with more accurate personalized recommendations through data mining and machine learning algorithms. Existing route recommendation systems have made significant progress in the following aspects: user portrait construction, multi-dimensional data fusion, intelligent algorithm optimization, real-time dynamic recommendations, and personalized customization services.

[0003] The recommendation results of existing travel route recommendation systems are too dependent on user historical data, have cold start problems for new users, and the recommended routes are too homogeneous. Summary of the invention

[0004] In order to overcome the problems of the prior art tourist route recommendation system that the recommendation results are too dependent on user historical data, cold start problems for new users and homogeneity of recommended routes, the present invention designs a tourist route recommendation system based on tourist consumer portraits and psychological models, which can effectively solve the technical problems of Shanshu.

[0005] In order to solve the above technical problems, the technical solution of the present invention is as follows:

[0006] A travel route recommendation system based on travel consumer portraits and psychological models, comprising: a data acquisition and preprocessing module, a user and scenic spot interaction information map construction module, a scenic spot knowledge map construction module, a collaborative knowledge map construction module, a route planning module, a deep learning model application module and a knowledge map application module;

[0007] The data acquisition and preprocessing module is used to acquire the user behavior data of tourism consumers from the tourism platform, and to clean and standardize the data. The output end of the data acquisition and preprocessing module is connected to the input end of the user and scenic spot interaction information graph construction module. The output end of the data acquisition and preprocessing module is also connected to the input end of the scenic spot knowledge graph construction module;

[0008] The user and scenic spot interaction information graph construction module is used to construct an interaction information graph between users and scenic spots based on user behavior data, and its output end is connected to the input end of the collaborative knowledge graph construction module;

[0009] The scenic spot knowledge graph construction module is used to construct the scenic spot knowledge graph using a top-down method and import it into the graph database, and its output end is connected to the input end of the collaborative knowledge graph construction module;

[0010] The collaborative knowledge graph construction module is used to combine user interaction information and scenic spot knowledge graph to construct a collaborative knowledge graph, and its output end is connected to the input end of the route planning module, and is also connected to the input end of the deep learning model application module, and is connected to the input end of the knowledge graph technology application module;

[0011] The route planning module is used to plan a tourist route according to the recommended attractions, and its output end is connected to the input end of the visualization module;

[0012] The deep learning model application module is used to introduce the GRU bidirectional neural network model to mine deep-level scenic spot attribute features and user interest features, and its output end is connected to the input end of the knowledge graph application module;

[0013] The knowledge graph application module is used to convert node parameters in the collaborative knowledge graph into vectors through the TransR model, and use the knowledge-aware attention mechanism to learn the weight of each neighbor during the propagation process, and its output end is connected to the input end of the route planning module;

[0014] The visualization module is used to display the travel route recommendation results;

[0015] Preferably, the user and scenic spot interaction information graph construction module includes:

[0016] A behavior data extraction unit, used to extract the interaction data between the user and the scenic spot from the user behavior log;

[0017] The interactive graph construction unit constructs an interactive information graph between the user and the scenic spot based on the interactive data using a graph data structure to reflect the user's preference and access mode for different scenic spots.

[0018] Preferably, the scenic spot knowledge graph construction module includes:

[0019] Knowledge graph extraction unit, used to extract knowledge information related to scenic spots from various channels;

[0020] The knowledge graph fusion unit integrates and verifies the extracted knowledge information to form a structured scenic spot knowledge base;

[0021] The graph database import unit imports the scenic spot knowledge base into the graph database and constructs the scenic spot knowledge graph for efficient query and reasoning.

[0022] Preferably, the collaborative knowledge graph construction module includes:

[0023] An information fusion unit, used to fuse the user interaction information graph and the scenic spot knowledge graph to form a collaborative knowledge graph containing user behavior and scenic spot knowledge;

[0024] The graph structure optimization unit performs structural optimization on the collaborative knowledge graph to improve the accuracy and usability of the graph.

[0025] Preferably, the route planning module includes:

[0026] The attraction recommendation unit recommends attractions of interest to users based on collaborative knowledge graphs and user portraits;

[0027] The route optimization unit optimizes and plans travel routes based on recommended attractions combined with geographic information and user preferences.

[0028] Preferably, the deep learning model application module includes:

[0029] The feature extraction unit uses the GRU bidirectional neural network model to extract deep features from user behavior data and scenic spot attribute data;

[0030] The interest matching unit performs interest matching between users and attractions based on the extracted features to improve the accuracy of recommendations.

[0031] Preferably, the knowledge graph application module includes:

[0032] A vector conversion unit, which converts node parameters in the collaborative knowledge graph into vector representations through a TransR model;

[0033] The weight learning unit uses the knowledge-aware attention mechanism to learn the weight of each neighbor node during the propagation process to optimize the recommendation results.

[0034] Compared with the prior art, the beneficial effects of the present invention are as follows: the system obtains user behavior data from the tourism platform through the data acquisition and preprocessing module, and cleans and standardizes it, providing a high-quality data foundation for subsequent modules. Then, the user-attraction interaction information graph construction module and the attraction knowledge graph construction module respectively construct the interaction information graph between users and attractions and the attraction knowledge graph, providing the system with rich user and attraction information. The collaborative knowledge graph construction module integrates the user interaction information and the attraction knowledge graph to form a collaborative knowledge graph containing user behavior and attraction knowledge, providing comprehensive information support for subsequent recommendations. On this basis, the route planning module plans the travel route according to the recommended attractions, combines geographic information and user preferences, and provides users with personalized travel routes. The deep learning model application module introduces the GRU bidirectional neural network model to mine deep-level attraction attribute features and user interest features, further improving the accuracy of the recommendation. At the same time, the knowledge graph application module converts the node parameters in the collaborative knowledge graph into vectors through the TransR model, and uses the knowledge-aware attention mechanism to learn the weight of each neighbor node, further optimizing the recommendation results. The visualization module displays the travel route recommendation results, so that users can intuitively understand the recommended travel routes. The entire system not only improves the accuracy and personalization of travel route recommendations, but also enhances the user experience through visual display. In addition, the various modules of the system work together to form a complete recommendation process, improve the efficiency and accuracy of recommendations, and provide users with a more convenient and efficient travel route recommendation service. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] In order to more clearly illustrate the implementation methods of the present invention or the technical solutions in the prior art, the drawings required for use in the implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings in the following description are only exemplary, and for ordinary technicians in this field, other implementation drawings can be derived based on the provided drawings without paying any creative work.

[0036] Figure 1 This is a structural diagram of a travel route recommendation system based on tourism consumer portraits and psychological models; DETAILED DESCRIPTION

[0037] The drawings are for illustrative purposes only and should not be construed as limiting the present patent;

[0038] In order to better illustrate the present embodiment, some parts in the drawings may be omitted, enlarged or reduced, and do not represent the size of the actual product;

[0039] It is understandable to those skilled in the art that some well-known structures and their descriptions may be omitted in the drawings.

[0040] The technical solution of the present invention is further described below in conjunction with the accompanying drawings and embodiments.

[0041] Example

[0042] A travel route recommendation system based on tourist consumer portraits and psychological models, such as Figure 1 As shown, it includes: data acquisition and preprocessing module, user and scenic spot interaction information graph construction module, scenic spot knowledge graph construction module, collaborative knowledge graph construction module, route planning module, deep learning model application module and knowledge graph application module;

[0043] The data acquisition and preprocessing module is used to acquire the user behavior data of tourism consumers from the tourism platform, and to clean and standardize the data. The output end of the data acquisition and preprocessing module is connected to the input end of the user and scenic spot interaction information graph construction module. The output end of the data acquisition and preprocessing module is also connected to the input end of the scenic spot knowledge graph construction module;

[0044] The user and scenic spot interaction information graph construction module is used to construct an interaction information graph between users and scenic spots based on user behavior data, and its output end is connected to the input end of the collaborative knowledge graph construction module;

[0045] The scenic spot knowledge graph construction module is used to construct the scenic spot knowledge graph using a top-down method and import it into the graph database, and its output end is connected to the input end of the collaborative knowledge graph construction module;

[0046] The collaborative knowledge graph construction module is used to combine user interaction information and scenic spot knowledge graph to construct a collaborative knowledge graph, and its output end is connected to the input end of the route planning module, and is also connected to the input end of the deep learning model application module, and is connected to the input end of the knowledge graph technology application module;

[0047] The route planning module is used to plan a tourist route according to the recommended attractions, and its output end is connected to the input end of the visualization module;

[0048] The deep learning model application module is used to introduce the GRU bidirectional neural network model to mine deep-level scenic spot attribute features and user interest features, and its output end is connected to the input end of the knowledge graph application module;

[0049] The knowledge graph application module is used to convert node parameters in the collaborative knowledge graph into vectors through the TransR model, and use the knowledge-aware attention mechanism to learn the weight of each neighbor during the propagation process, and its output end is connected to the input end of the route planning module;

[0050] The visualization module is used to display the travel route recommendation results;

[0051] The user and scenic spot interaction information graph construction module includes:

[0052] A behavior data extraction unit, used to extract the interaction data between the user and the scenic spot from the user behavior log;

[0053] First, obtain user behavior logs from the database of the tourism platform. These logs may exist in the form of text, database records or log files. Clean the collected logs to remove invalid, duplicate or incorrectly formatted records. Ensure that each record in the log contains key information such as user ID, attraction ID, behavior type (such as browsing, purchasing, etc.) and timestamp. Extract the interaction data between users and attractions based on the cleaned logs, which includes the number of times users visit attractions, length of stay, behavior type (such as browsing, evaluation, etc.) and time when the behavior occurs. Preprocess the extracted data, such as calculating the average length of stay of users at each attraction and counting the total number of visits of users, so that they can be used in the subsequent construction of interactive information graphs.

[0054] The interactive graph construction unit constructs an interactive information graph between the user and the scenic spot based on the interactive data using a graph data structure to reflect the user's preference and access mode for different scenic spots.

[0055] In the graph data structure, two types of nodes are defined: user nodes and attraction nodes. Each user node represents a unique user, and each attraction node represents a unique attraction.

[0056] Edge definition, establish edges between user nodes and attraction nodes to represent the interaction relationship between users and attractions. Each edge can contain multiple attributes, such as number of interactions, average stay time, behavior type, etc.

[0057] Graph construction,Based on the extracted interaction data, an interaction information graph between,users and attractions is constructed, which usually involves the following steps:

[0058] Create an empty graph data structure;

[0059] Traverse the extracted interaction data and create a user node for each user and a scenic spot node for each scenic spot if they do not already exist in the graph);

[0060] According to the interaction data between users and scenic spots, edges are established between user nodes and scenic spot nodes, and corresponding attributes are set.

[0061] Graph optimization: Optimize the constructed interaction information graph to improve the performance of subsequent algorithms. This may include merging similar nodes, deleting unimportant edges, etc.

[0062] Graph storage, stores the constructed interactive information graph in an appropriate graph database or file for use by subsequent modules.

[0063] The scenic spot knowledge graph construction module includes:

[0064] The knowledge graph extraction unit is used to extract knowledge information related to scenic spots from various channels, such as online resources and travel guides;

[0065] Select the most relevant and reliable data source based on the characteristics and needs of the attraction. For example, for historical and cultural heritage attractions, you may focus more on historical documents and official travel guides; while for natural scenery attractions, you may rely more on online resources and tourist reviews, and use natural language processing (NLP) and machine learning technology to extract key information from the selected data source. This information may include the name, location, historical background, opening hours, ticket price, transportation methods, surrounding facilities, etc. of the attraction. Clean the extracted information to remove duplicate, invalid or incorrectly formatted data, and at the same time, ensure the accuracy and consistency of the information.

[0066] The knowledge graph fusion unit integrates and verifies the extracted knowledge information to form a structured scenic spot knowledge base;

[0067] Integrate information from different data sources to eliminate information redundancy and conflict, which may involve steps such as data deduplication, entity alignment and attribute merging. Verify the integrated knowledge to ensure the accuracy and reliability of the information, which may include comparing with known facts, verifying with expert systems, etc. Represent the verified knowledge in a structured manner to form a scenic spot knowledge base, which may involve defining concepts such as entities, attributes and relationships, and building corresponding data models.

[0068] The graph database import unit imports the scenic spot knowledge base into the graph database and constructs the scenic spot knowledge graph for efficient query and reasoning.

[0069] According to the characteristics and requirements of the knowledge graph, select a suitable graph database. The graph database should support efficient query and reasoning operations and have good scalability and stability. Import the data in the scenic spot knowledge base into the graph database, which may involve steps such as data conversion and format adjustment to ensure that the data can be correctly stored in the graph database. Build a scenic spot knowledge graph in the graph database, which includes defining elements such as entity nodes, relationship edges and attributes, and building corresponding indexes and query mechanisms. Optimize the performance of the constructed knowledge graph to improve the efficiency of query and reasoning, which may include index optimization, cache strategy adjustment, etc.

[0070] The collaborative knowledge graph construction module includes:

[0071] An information fusion unit, used to fuse the user interaction information graph and the scenic spot knowledge graph to form a collaborative knowledge graph containing user behavior and scenic spot knowledge;

[0072] Determine the key points of fusion, that is, how to effectively combine user behavior and scenic spot knowledge, adopt appropriate technologies and methods, such as entity alignment, relationship mapping, etc., match and associate the entities and relationships in the user interaction information graph with the corresponding parts in the scenic spot knowledge graph, perform entity alignment, ensure that the entities in the user interaction information graph and the entities in the scenic spot knowledge graph can accurately correspond, perform relationship mapping, associate the interaction relationship between users and attractions with the relationship in the scenic spot knowledge graph, form a richer relationship network, integrate attributes and information, integrate the relevant attributes of user behavior and scenic spot knowledge, and form a complete entity description.

[0073] The graph structure optimization unit performs structural optimization on the collaborative knowledge graph to improve the accuracy and usability of the graph.

[0074] Identify and remove redundant information in the graph, such as duplicate entities, relationships, or attributes.

[0075] Ensure that the information in the graph is concise and clear to facilitate subsequent analysis and application, analyze the relationship network in the graph, and identify possible missing relationships.

[0076] Based on known information and logical rules, missing relationships are completed to make the map more complete and accurate, and the importance of entities is evaluated based on users' interactive behaviors at scenic spots.

[0077] Important entities are marked or weighted so that they can be given more attention in subsequent analysis.

[0078] The route planning module includes:

[0079] The attraction recommendation unit recommends attractions of interest to users based on collaborative knowledge graphs and user portraits;

[0080] Build user portraits from multi-dimensional information such as user historical behavior, preferences, interests, etc.

[0081] User portraits may include the user's age, gender, occupation, travel preferences such as natural scenery, historical culture, food experience, travel time, budget, etc., and use the rich information in the collaborative knowledge graph, including the attributes of attractions such as type, location, opening hours, ticket prices, etc., user-attraction interaction data such as number of visits, length of stay, evaluations, etc., as well as relationships between attractions, such as adjacent relationships, similarities, etc.

[0082] By analyzing the information in the user portrait and the collaborative knowledge graph, we identify attractions that match the user's interests. We use collaborative filtering, content recommendation, graph-based recommendation and other algorithms to generate a personalized attraction recommendation list for the user based on the user portrait and the information in the collaborative knowledge graph. The recommended attractions may be sorted according to the user's preferences, such as the degree of interest the user is most likely to have, the attraction's evaluation score, and the distance from the user's current location or planned travel destination.

[0083] The route optimization unit optimizes and plans travel routes based on recommended attractions, combined with geographic information and user preferences.

[0084] Collect and integrate geographic information related to scenic spots, such as the geographical location of scenic spots, transportation methods, such as bus, subway, self-driving, etc., travel time, distance, etc., ensure the accuracy and real-time nature of geographic information, so as to carry out accurate route planning, analyze the preference information in user portraits, such as whether users prefer low-carbon travel methods such as walking and cycling, whether they want to avoid traffic congestion during peak hours, whether they want to arrange enough rest time, etc., integrate user preferences into route planning to improve users' travel experience, adopt path planning algorithms, such as Dijkstra algorithm, A* algorithm, genetic algorithm, etc., generate multiple possible travel routes for users based on recommended scenic spots and geographic information, consider factors such as the opening hours of scenic spots, user travel time, traffic time, etc., screen and optimize routes, evaluate the feasibility, time consumption, cost, etc. of each route, recommend the best travel route for users, and present the planned travel route in a graphical way, such as map, route map, etc. In the visual interface, users can intuitively see the location of scenic spots, transportation methods, estimated time consumption and other information, which is convenient for itinerary arrangement and adjustment.

[0085] The deep learning model application module includes:

[0086] The feature extraction unit uses the GRU bidirectional neural network model to extract deep features from user behavior data and scenic spot attribute data;

[0087] GRU was chosen as a variant of recurrent neural network (RNN) because it can effectively handle long-term dependencies in sequential data.

[0088] A bidirectional GRU structure is adopted, which considers both forward and reverse sequence information at the same time to capture more comprehensive contextual features.

[0089] User behavior data and scenic spot attribute data are input into the bidirectional GRU model respectively.

[0090] The model learns deep feature representations in the data through forward propagation and back propagation.

[0091] Feature vectors of users and attractions are extracted, which contain rich semantic information and contextual associations.

[0092] The interest matching unit performs interest matching between users and attractions based on the extracted features to improve the accuracy of recommendations.

[0093] The similarity between the user feature vector and the scenic spot feature vector is calculated using measurement methods such as cosine similarity and Euclidean distance.

[0094] The higher the similarity, the higher the interest match between the user and the attraction.

[0095] Based on the calculated similarity, an interest matching algorithm is designed to recommend attractions that best match users' interests.

[0096] The algorithm may consider multiple factors, such as user preferences, the type of attraction, the interaction history between the user and the attraction, etc.

[0097] Generate a personalized list of recommended attractions based on the results of the interest matching algorithm.

[0098] The recommendation list may be sorted according to the user's preferences, such as by similarity from high to low, the evaluation score of the attraction, the distance from the user's current location or planned travel destination, etc.

[0099] The knowledge graph application module includes:

[0100] A vector conversion unit, which converts node parameters in the collaborative knowledge graph into vector representations through a TransR model;

[0101] First, all nodes and relationships are extracted from the collaborative knowledge graph to form a training data set. Then, the TransR model is used to train the training data set. During the training process, the model will learn the vector representation of each node and the vector representation of the relationship between nodes. These vector representations will capture the semantic information of the nodes and relationships. After training, the model will output the vector representation of each node. These vector representations can be used in subsequent recommendation algorithms.

[0102] The weight learning unit uses the knowledge-aware attention mechanism to learn the weight of each neighbor node during the propagation process to optimize the recommendation results.

[0103] First, for a given target node (such as a user or a scenic spot), its neighbor nodes are selected from the collaborative knowledge graph. These neighbor nodes may be nodes that have a direct relationship with the target node or nodes that are reachable through some path.

[0104] Then, the knowledge-aware attention mechanism is used to calculate the weight of each neighbor node. The attention mechanism calculates the weight based on the vector representation of the neighbor nodes and the vector representation of the relationship between them and the target node. The size of the weight reflects the importance of the neighbor node to the target node.

[0105] Finally, the calculated weights are applied to the vector representations of the neighboring nodes and weighted summed up. In this way, a new vector representation of the target node can be obtained, which incorporates the information of its neighboring nodes.

[0106] Using this new vector representation, a recommendation algorithm can be further applied to recommend attractions or travel routes to users. The recommendation algorithm can make recommendations based on the similarity between the user vector and the attraction vector, or based on the combined similarity between the user vector and the attraction vectors in the travel route.

[0107] In the specific implementation, the data acquisition and preprocessing module first obtained a large amount of user behavior data of travel consumers from a well-known travel platform. These data included users' browsing records, purchase records, evaluation records, etc. The module cleaned and standardized these data, removed duplicate data and invalid data, and unified the data format to ensure that subsequent modules can process these data accurately and efficiently.

[0108] Next, the user-attraction interaction information graph construction module extracted the user-attraction interaction data from the cleaned data, including the number of times users visited the attractions, the length of stay, and evaluations. Then, the module used the graph data structure to construct an interaction information graph between users and attractions, which clearly reflects users' preferences and visit patterns for different attractions.

[0109] At the same time, the scenic spot knowledge graph construction module extracts knowledge information related to the scenic spot from online resources, travel guides and other channels, such as the historical background, geographical location, transportation methods, opening hours, etc. of the scenic spot. The module integrates and verifies this knowledge information to form a structured scenic spot knowledge base, and imports it into the graph database to construct a scenic spot knowledge graph.

[0110] Then, the collaborative knowledge graph construction module integrates the user interaction information graph and the scenic spot knowledge graph to form a collaborative knowledge graph that includes user behavior and scenic spot knowledge. The module also optimizes the structure of the collaborative knowledge graph to improve the accuracy and usability of the graph.

[0111] Based on the collaborative knowledge graph, the route planning module recommends attractions of interest to users, and optimizes and plans travel routes based on the recommended attractions, combined with geographic information and user preferences. For example, for users who like natural scenery and prefer self-driving tours, the module plans a self-driving route for them that includes multiple famous natural landscapes.

[0112] At the same time, the deep learning model application module introduced the GRU bidirectional neural network model, extracted deep features from user behavior data and attraction attribute data, and performed interest matching between users and attractions, which further improved the accuracy of recommendations and enabled the system to recommend attractions that are more in line with users' interests and needs.

[0113] In addition, the knowledge graph application module converts the node parameters in the collaborative knowledge graph into vector representations through the TransR model, and uses the knowledge-aware attention mechanism to learn the weight of each neighbor node during the propagation process, which helps to optimize the recommendation results and enables the system to recommend more suitable attractions and travel routes for users.

[0114] Finally, the visualization module displays the recommended travel route results. Users can intuitively see the recommended travel routes and attractions information through the system interface, which not only improves the user experience, but also makes the system more practical and convenient.

[0115] The same or similar reference numerals correspond to the same or similar components;

[0116] The terms used in the drawings to describe positional relationships are only used for illustrative purposes and should not be construed as limiting this patent;

[0117] Obviously, the above embodiments of the present invention are merely examples for clearly illustrating the present invention, and are not limitations on the implementation methods of the present invention. For ordinary technicians in the relevant field, other different forms of changes or modifications can be made on the basis of the above description. It is not necessary and impossible to list all the implementation methods here. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the claims of the present invention.

Claims

1. A travel route recommendation system based on tourist consumer portraits and psychological models, characterized in that: include: Data acquisition and preprocessing module, user and scenic spot interaction information graph construction module, scenic spot knowledge graph construction module, collaborative knowledge graph construction module, route planning module, deep learning model application module, knowledge graph application module and visualization module; The data acquisition and preprocessing module is used to acquire the user behavior data of tourism consumers from the tourism platform, and to clean and standardize the data. The output end of the data acquisition and preprocessing module is connected to the input end of the user and scenic spot interaction information graph construction module. The output end of the data acquisition and preprocessing module is also connected to the input end of the scenic spot knowledge graph construction module; The user and scenic spot interaction information graph construction module is used to construct an interaction information graph between users and scenic spots based on user behavior data, and its output end is connected to the input end of the collaborative knowledge graph construction module; The scenic spot knowledge graph construction module is used to construct the scenic spot knowledge graph using a top-down method and import it into the graph database, and its output end is connected to the input end of the collaborative knowledge graph construction module; The collaborative knowledge graph construction module is used to combine user interaction information and scenic spot knowledge graph to construct a collaborative knowledge graph, and its output end is connected to the input end of the route planning module, and is also connected to the input end of the deep learning model application module, and is connected to the input end of the knowledge graph technology application module; The route planning module is used to plan a tourist route according to the recommended attractions, and its output end is connected to the input end of the visualization module; The deep learning model application module is used to introduce the GRU bidirectional neural network model to mine deep-level scenic spot attribute features and user interest features, and its output end is connected to the input end of the knowledge graph application module; The knowledge graph application module is used to convert node parameters in the collaborative knowledge graph into vectors through the TransR model, and use the knowledge-aware attention mechanism to learn the weight of each neighbor during the propagation process, and its output end is connected to the input end of the route planning module; The visualization module is used to display the travel route recommendation results.

2. A travel route recommendation system based on travel consumer portraits and psychological models according to claim 1, characterized in that: The user and scenic spot interaction information graph construction module includes: A behavior data extraction unit, used to extract the interaction data between the user and the scenic spot from the user behavior log; The interactive graph construction unit constructs an interactive information graph between the user and the scenic spot based on the interactive data using a graph data structure to reflect the user's preference and access mode for different scenic spots.

3. A travel route recommendation system based on travel consumer portraits and psychological models according to claim 1, characterized in that: The scenic spot knowledge graph construction module includes: The knowledge graph extraction unit is used to extract knowledge information related to scenic spots from various channels (online resources, travel guides); The knowledge graph fusion unit integrates and verifies the extracted knowledge information to form a structured scenic spot knowledge base; The graph database import unit imports the scenic spot knowledge base into the graph database and constructs the scenic spot knowledge graph for efficient query and reasoning.

4. A travel route recommendation system based on travel consumer portraits and psychological models according to claim 1, characterized in that: The collaborative knowledge graph construction module includes: An information fusion unit, used to fuse the user interaction information graph and the scenic spot knowledge graph to form a collaborative knowledge graph containing user behavior and scenic spot knowledge; The graph structure optimization unit performs structural optimization on the collaborative knowledge graph to improve the accuracy and usability of the graph.

5. A travel route recommendation system based on travel consumer portraits and psychological models according to claim 1, characterized in that: The route planning module includes: The attraction recommendation unit recommends attractions of interest to users based on collaborative knowledge graphs and user portraits; The route optimization unit optimizes and plans travel routes based on recommended attractions combined with geographic information and user preferences.

6. A travel route recommendation system based on travel consumer portraits and psychological models according to claim 1, characterized in that: The deep learning model application module includes: The feature extraction unit uses the GRU bidirectional neural network model to extract deep features from user behavior data and scenic spot attribute data; The interest matching unit performs interest matching between users and attractions based on the extracted features to improve the accuracy of recommendations.

7. A travel route recommendation system based on travel consumer portraits and psychological models according to claim 1, characterized in that: The knowledge graph application module includes: A vector conversion unit, which converts node parameters in the collaborative knowledge graph into vector representations through a TransR model; The weight learning unit uses the knowledge-aware attention mechanism to learn the weight of each neighbor node during the propagation process to optimize the recommendation results.

Images