Route generation method and device, electronic equipment, storage medium and computer program product

Abstract

The invention provides a route generation method and device, electronic equipment, a storage medium and a computer program product, and relates to the technical field of artificial intelligence, and the method comprises the following steps: in the embodiment of the invention, performing data extraction on acquired tourism data by using a large model, and determining a plurality of routes; and determining a target route based on the similarity between the related data corresponding to the plurality of routes and the demand data corresponding to the current user. Through the scheme in the embodiment of the invention, a personalized target route can be provided.

Description

Technical Field

[0001] This application relates to the field of artificial intelligence technology, and in particular to a route generation method, apparatus, electronic device, storage medium, and computer program product. Background Technology

[0002] Nowadays, travel routes need to be planned in advance, and people often get travel routes from the internet. However, the quality of online route recommendations varies, and they are not always the same as people's actual experiences, failing to meet users' personalized needs for travel routes. Summary of the Invention

[0003] The embodiments of this application provide a route generation method, apparatus, electronic device, storage medium, and computer program product that can provide users with personalized travel routes.

[0004] The technical solution of this application is implemented as follows:

[0005] This application provides a route generation method, including:

[0006] The acquired tourism data is used to extract data using a large model to determine multiple routes;

[0007] The target route is determined based on the similarity between the relevant data corresponding to the multiple routes and the current user's demand data.

[0008] In the above scheme, the process of extracting data from the acquired tourism data using a large model to determine multiple routes includes:

[0009] Using the large model, extract text data in a uniform format from the acquired tourism data;

[0010] Extract multiple key entities from the text data and establish the relationships between these key entities.

[0011] Among the multiple key entities with related relationships, multiple routes are extracted; wherein the related data of the routes includes at least one of the following: node-related entities, time-related entities, object-related entities, evaluation-related entities, and route mode entities.

[0012] In the above scheme, the relevant data further includes: route features; after extracting data from the acquired tourism data using a large model to determine multiple routes, before determining the target route based on the similarity between the relevant data corresponding to each of the multiple routes and the current user's corresponding demand data, the method further includes:

[0013] Determine the route features corresponding to each of the multiple routes.

[0014] In the above scheme, the tourism data includes: basic data from multiple nodes and travelogue data corresponding to multiple users; the step of using the large model to extract text data in a unified format from the acquired tourism data includes:

[0015] Using the large model, the acquired basic data of multiple nodes and the travelogue data corresponding to multiple users are processed by text conversion to obtain initial text data;

[0016] The initial text data is standardized and filtered to extract text data in a uniform format.

[0017] In the above scheme, after extracting uniformly formatted text data from the acquired tourism data using the large model, and before extracting multiple key entities from the text data and establishing the relevant relationships between the key entities, the method further includes:

[0018] The text content in the text data is standardized and adjusted using a preset language model.

[0019] In the above scheme, the step of extracting multiple key entities from the text data and establishing the relevant relationships between the key entities includes:

[0020] Multiple key entities are extracted from the text data based on a preset entity template; wherein, the preset entity template includes: preset node-related entities, preset time-related entities, preset object-related entities, and preset evaluation entities;

[0021] A correlation analysis is performed on multiple key entities to establish the relationships between them.

[0022] In the above scheme, after extracting multiple routes from the multiple key entities with relevant relationships, and before determining the route features corresponding to each of the multiple routes, the method further includes:

[0023] Based on the similarity between each of the routes, routes with a similarity greater than a preset threshold are deleted.

[0024] In the above scheme, determining the route features corresponding to the multiple routes includes:

[0025] For each of the multiple routes, the node distribution characteristics corresponding to each route are determined based on the geographical location features and node type features of the node-related entities represented by each route.

[0026] Based on the node identifier of the node-related entity representation corresponding to each route, and the similarity between the node identifier of the node-related entity representation corresponding to other routes, the unique feature corresponding to each route is determined.

[0027] Based on the evaluation information of each user represented by the evaluation-related entity corresponding to each route, the evaluation features corresponding to each route are determined.

[0028] The node distribution characteristics, the uniqueness characteristics, and the evaluation characteristics are weighted and summed to determine each route characteristic.

[0029] In the above scheme, determining the target route based on the similarity between the relevant data corresponding to the multiple routes and the current user's demand data includes:

[0030] The entity corresponding to each route is processed using a preset conversion model, and the corresponding route features are used to determine the corresponding first preset format features. The demand data is then processed using the preset conversion model to determine the corresponding second preset format features.

[0031] Retrieve from multiple first preset format features the top N first preset format features that have the highest similarity to the second preset format features; N is an integer greater than 0;

[0032] The target route is determined based on the N routes and the demand data corresponding to the first N first preset format features.

[0033] This application also provides a route generation apparatus, including:

[0034] The extraction unit is used to extract data from the acquired tourism data using a large model and determine multiple routes;

[0035] The determining unit is used to determine the target route based on the similarity between the multiple routes and the current user's corresponding demand data.

[0036] This application also provides an electronic device, including a memory and a processor. The memory stores a computer program that can run on the processor, and the processor executes the computer program to implement the steps in the above-described method.

[0037] This application also provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the steps in the above-described method.

[0038] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the steps in the above-described method.

[0039] In this embodiment, a large-scale model is used to extract data from the acquired tourism data to determine multiple route features. Based on the similarity between the relevant data corresponding to each of the multiple routes and the current user's demand data, a target route is determined. This allows for efficient and accurate analysis of the acquired tourism data using a large-scale model, extracting multiple routes with significant characteristics. Then, by comparing the similarity of these significant routes with the current user's demand data, routes that adaptively determine individual needs, while considering diversity and practicality, can be provided, thus offering personalized target routes. Attached Figure Description

[0040] Figure 1 An optional flowchart illustrating the route generation method provided in an embodiment of this application;

[0041] Figure 2 An optional flowchart illustrating the route generation method provided in an embodiment of this application;

[0042] Figure 3 An optional flowchart illustrating the route generation method provided in an embodiment of this application;

[0043] Figure 4 An optional flowchart illustrating the route generation method provided in an embodiment of this application;

[0044] Figure 5 An optional flowchart illustrating the route generation method provided in an embodiment of this application;

[0045] Figure 6 An optional flowchart illustrating the route generation method provided in an embodiment of this application;

[0046] Figure 7 This is a schematic diagram of the route generation device provided in the embodiments of this application;

[0047] Figure 8 This is a schematic diagram of a hardware entity of an electronic device provided in an embodiment of this application. Detailed Implementation

[0048] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application are further described in detail below with reference to the accompanying drawings and embodiments. The described embodiments should not be regarded as limitations on this application. All other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0049] In the following description, references are made to “some embodiments,” which describe a subset of all possible embodiments. However, it is understood that “some embodiments” may be the same subset or different subsets of all possible embodiments and may be combined with each other without conflict.

[0050] If the application documents contain similar descriptions such as "first / second", the following explanation shall be added: In the following description, the terms "first / second / third" are used only to distinguish similar objects and do not represent a specific order of objects. It is understood that "first / second / third" may be interchanged in a specific order or sequence where permitted, so that the embodiments of this application described herein can be implemented in an order other than that illustrated or described herein.

[0051] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of this application only and is not intended to limit this application.

[0052] This application provides a route generation method. Please refer to [link / reference]. Figure 1 This is an optional flowchart illustrating the route generation method provided in this application embodiment, which will be combined with... Figure 1 The steps shown are explained below:

[0053] S101. Use a large model to extract data from the acquired tourism data and determine multiple routes.

[0054] In this embodiment of the application, the route generation device can acquire a large amount of tourism-related data from the entire network and use a large model to extract data from the tourism-related data, including multiple routes.

[0055] In this embodiment, the route generation device can utilize large models such as deep learning models, natural language processing models, and geographic information systems to efficiently and accurately parse tourism data and extract multiple routes. These large models can include any of the following: convolutional neural network (CNN) models, deep belief network (DBN) models, stacked auto-encoder network (SAEN) models, natural language processing (NLP) models, Transformer models, etc.

[0056] The tourism data can include basic data on various attractions (nodes) in the network, as well as travelogue data posted by multiple users. Routes can include route node data, time-related data, object-related data, evaluation-related data, and route mode data recorded in the tourism data. The route generation device can be a server, cloud server, or terminal with corresponding data processing capabilities.

[0057] S102. Determine the target route based on the similarity between the relevant data corresponding to the multiple routes and the current user's demand data.

[0058] In this embodiment, the route generation device can acquire the current user's travel demand data based on a human-computer interaction device. It extracts corresponding features from this demand data and determines the relevant data corresponding to multiple routes. Based on the similarity between the features of the demand data and the relevant data corresponding to the multiple routes, it determines the top N routes with the highest similarity. The target route is then determined based on the top N routes and the demand data. N is an integer greater than 0. The relevant data may include route features and text entities.

[0059] In this embodiment of the application, the route generation device can also convert the demand data and the relevant data of the route into feature information in the same format, and perform similarity retrieval based on the feature information in the same format to determine the top N routes with the highest similarity.

[0060] In this embodiment, a large-scale model is used to extract data from the acquired tourism data to determine multiple routes. Based on the similarity between the relevant data corresponding to each of the multiple routes and the current user's demand data, a target route is determined. This allows for efficient and accurate analysis of the acquired tourism data using a large-scale model, extracting multiple routes with significant similarities. Furthermore, comparing the similarity of these significant routes with the current user's demand data enables adaptive determination of a tourism plan that meets individual needs, balances diversity and practicality, and ultimately provides a personalized target route.

[0061] Please see Figure 2 This is an optional flowchart illustrating the route generation method provided in an embodiment of this application. Figure 1 S101 shown can also be implemented by S201 to S204, which will be explained in conjunction with the steps:

[0062] S201. Using the large model, extract text data in a unified format from the acquired tourism data.

[0063] In this embodiment of the application, the route generation device can use a large model to convert tourism data into initial text data, and preprocess and purify the initial text data to obtain text data in a unified format.

[0064] S202. Extract multiple key entities from the text data and establish the correlation between the key entities.

[0065] In this embodiment of the application, the route generation device can also extract multiple key entities from the text data, and use a joint extraction model to identify the relationship between each key entity and establish the correlation between the relevant key entities.

[0066] Key entities may include: tourist attractions (nodes), tourist time, attractions features, attractions evaluation, routes and modes of travel, and tourist objects.

[0067] S203. Extract multiple routes from the multiple key entities that have a relationship with each other.

[0068] In this embodiment, the route generation device can extract multiple routes from multiple key entities with related relationships using a first-class model trained on a Chinese fine-tuned version of LLAMA3. The related data for the routes includes at least one of the following: node-related entities, time-related entities, object-related entities, evaluation-related entities, and route mode entities.

[0069] In this embodiment of the application, the route generation device can input multiple key entities with related relationships into the first large model trained based on the Chinese fine-tuning version of LLAMA3 to extract the route. The route example is "Please extract the number of travel days, number of people, travel mode (self-driving, group tour), Day 1 time, location, number of people, and attraction name and introduction based on the input entities and travelogue information".

[0070] In this embodiment of the application, after extracting multiple routes, the route generation device can also delete routes whose similarity is greater than a preset threshold based on the similarity between each route.

[0071] For example, the route generation device can generate a hash signature for each extracted route using the MinHASH algorithm. This hash signature reflects the core characteristics of the route. Then, Locality-Sensitive Hashing (LSH) technology is implemented to quickly identify and remove duplicate or highly similar routes. The specific similarity removal formula is as follows: MinHash(route) → Hashk, if d(Hashi, Hashj) < θ, then duplicates are removed. Here, d represents the Euclidean distance between Hashi and Hashj.

[0072] S204. Determine the route features corresponding to each of the multiple routes.

[0073] In this embodiment, the route generation device can determine the node distribution characteristics, uniqueness characteristics, and evaluation characteristics of each route based on entity analysis of the multiple routes. Based on the node distribution characteristics, uniqueness characteristics, and evaluation characteristics of each route, corresponding route characteristics are determined, thereby obtaining multiple route characteristics.

[0074] In this embodiment, a large amount of tourism data is obtained from the large model hedging network and processed in multiple stages. This allows for the extraction of numerous real routes from a vast amount of real tourism data, solving the problem of high error rates in route generation using large models in related technologies and thus avoiding the "large model illusion." Furthermore, the determined route features include three dimensions of characteristics corresponding to the route, providing a more comprehensive consideration and resulting in higher significance of the determined route features. Therefore, the target routes determined using significant route features and user demand data are more accurate and personalized.

[0075] Please see Figure 3 This is an optional flowchart illustrating the route generation method provided in an embodiment of this application. Figure 2 S201 shown can also be implemented via S301 to S302, which will be explained in conjunction with the steps:

[0076] S301. Using the large model, perform text conversion processing on the acquired basic data of multiple nodes and the travelogue data corresponding to multiple users respectively to obtain initial text data.

[0077] In this embodiment, the route generation device can acquire basic data of multiple nodes and travelogue data corresponding to multiple users on the network. Since the acquired data comes in various formats, text conversion processing can be performed on the acquired location information and travelogue data to obtain initial text data.

[0078] In this embodiment, the route generation device comprehensively collects basic information (node ​​basic information) of major tourist attractions, including attraction introductions, details of supporting facilities, opening and closing times, etc., laying a data foundation for subsequent personalized planning. It also utilizes web crawling technology or data interface integration to extensively crawl travelogue data published by users from various online platforms, forming a vast amount of travelogue data knowledge, providing rich sources of material for subsequent analysis.

[0079] S302. Standardize and filter the initial text data to extract text data in a uniform format.

[0080] In this embodiment of the application, the route generation device can perform noise removal, text standardization and filtering on the initial text data, and then extract text data in a uniform format from the initial text data.

[0081] In this embodiment, noise removal involves using advanced regular expressions and machine learning models to identify and filter out invalid information such as advertising content, spam comments, and special characters from the initial text data. Text standardization includes standardizing capitalization and punctuation in the initial text data, as well as standardizing region-specific expressions to ensure text consistency across geographical regions. Filtering utilizes language recognition technology to exclude non-target language text (non-Chinese) from the initial text data, focusing on improving the relevance and purity of the dataset.

[0082] In this embodiment of the application, after extracting the text data, the text content in the text data can be standardized and adjusted using a preset language model. For example, the Bidirectional Encoder Representations from Transformers (BERT) model can be used to fine-tune the model for tourism-related vocabulary and expressions in the text data, thereby enhancing the ability to understand context.

[0083] In this embodiment, the large model is used to perform text conversion processing on the acquired basic data of multiple nodes and the travelogue data corresponding to multiple users to obtain initial text data. The initial text data is then standardized and filtered to extract text data in a uniform format. This standardization and filtering of the initial text data effectively removes noise, generating routes with fewer noisy entities—that is, more accurate routes. Furthermore, the target routes determined by combining the route features of these more accurate routes with user demand data are more accurate and personalized.

[0084] Please see Figure 4 This is an optional flowchart illustrating the route generation method provided in an embodiment of this application. Figure 2 S202 shown can also be implemented via S401 to S402, which will be explained in conjunction with the steps:

[0085] S401. Extract multiple key entities from the text data based on a preset entity template.

[0086] In this embodiment of the application, the route generation device can extract multiple key entities from text data based on the Named Entity Recognition (NER) model and with reference to a preset entity template.

[0087] The preset entity template includes: preset node-related entities, preset time-related entities, preset object-related entities, and preset evaluation entities.

[0088] In this embodiment, the route generation device can perform entity extraction (SemanticEncodingsNER(Entities,Types)={(e1,t1),(e2,t2),…}) on text data. This accurately identifies and classifies key entities in the text data, such as location entities (attraction names, city names), time entities (travel dates, time periods), evaluation entities (positive / negative sentiment), and characteristic entities (natural landscapes, historical sites), etc.

[0089] S402. Perform correlation analysis on multiple key entities to establish the correlation relationships between the key entities.

[0090] In this embodiment of the application, the route generation device can perform correlation analysis on multiple key entities based on a joint extraction model, determine the correlation between at least two key entities, and establish the correlation between key entities with correlation.

[0091] Establishing relationships between key entities with related relationships can include: storing key entities with related relationships in a set, or configuring corresponding relationship identifiers for key entities with related relationships.

[0092] For example, the route generation device can use a joint extraction model to identify complex relationships between key entities, such as "tour time-attraction" and "evaluation-activity".

[0093] In this embodiment, multiple key entities are extracted from the text data based on a preset entity template. Correlation analysis is then performed on these key entities to establish relationships between them. This establishes the relationships between the key entities, facilitating the joint extraction model to extract accurate routes from among them and reducing the generation of erroneous routes.

[0094] Please see Figure 4 This is an optional flowchart illustrating the route generation method provided in an embodiment of this application. Figure 2 S204 shown can also be implemented via S501 to S504, which will be explained in conjunction with the steps:

[0095] S501. For each of the multiple routes, based on the geographical location features and node type features of the node-related entities corresponding to each route, determine the node distribution features corresponding to each route.

[0096] In this embodiment, since the node-related entities included in each route can characterize the geographical location features and node type features of the nodes included in that route, for each of multiple routes, the distance between every two nodes can be determined based on the location features corresponding to the nodes in that route, and the node distribution features corresponding to that route can be determined based on the distance and the node type features. The geographical location features can include latitude and longitude information, and the node type features can include the travel type corresponding to the node.

[0097] For example, for a node in each route, the corresponding node distribution feature A can be determined using formula (1), and then the node distribution feature corresponding to the route can be determined based on the node distribution feature corresponding to each node.

[0098]

[0099] Where K is the number of nodes in the route, dist(i,j) is the distance between node i and node j, T is a hyperparameter, and unique_types is the uniqueness determined based on node type features.

[0100] S502. Based on the similarity between the node identifier of the node-related entity representation corresponding to each route and the node identifier of the node-related entity representation corresponding to other routes, determine the unique feature corresponding to each route.

[0101] In this embodiment of the application, since the node-related entities included in each route may include the node identifiers included in that route, a first set of similar node identifiers can be determined based on the similarity between the node identifiers in this route and the node identifiers in other routes. The number of the first set of node identifiers is then divided by the number of nodes in the union of the node identifiers in this route and the node identifiers in other routes to obtain a ratio. The uniqueness feature is determined based on this ratio.

[0102] For example, the Jaccard similarity coefficient is used to measure the similarity to other routes: Here, the Jaccard similarity coefficient is used to measure the similarity to other routes. Ri is the set of node identifiers corresponding to this route, and Rj is the set of node identifiers for other routes. The uniqueness feature B can be calculated using formula (2):

[0103]

[0104] S503. Based on the evaluation information corresponding to each user represented by the evaluation-related entity corresponding to each route, determine the evaluation features corresponding to each route.

[0105] In this embodiment, since each route includes evaluation entities that can include evaluations from multiple users for that route, the route generation device can determine the evaluation features corresponding to each route based on the ratings determined from the evaluation information and the sentiment tendency of the comments corresponding to each evaluation.

[0106] For example, the evaluation feature C can be calculated using formula (3):

[0107]

[0108] Where M is the number of users who rated the route, Si is the rating of the i-th user, and sentiment(R) i The parameter is determined based on the emotional positivity of the i-th user's evaluation.

[0109] S504. The node distribution characteristics, the uniqueness characteristics, and the evaluation characteristics are weighted and summed to determine each of the route characteristics.

[0110] In this embodiment, the route generation device can determine each of the route features by performing a weighted summation based on the node distribution features, the uniqueness features, and the evaluation features.

[0111] For example, the route feature score for each route can be determined using formula (4):

[0112] Score=ω A ·A+ω B ·B+ω C ·C (4)

[0113] Where, ω A Let ω be the weight corresponding to A. B ω represents the weight corresponding to B. C The weight corresponding to C.

[0114] In this embodiment, the corresponding node distribution characteristics, uniqueness characteristics, and evaluation characteristics are determined based on the entities included in the route. Then, a weighted sum of the node distribution characteristics, uniqueness characteristics, and evaluation characteristics is performed to determine each route characteristic. This approach considers three dimensions of features during the route characteristic determination process, providing a more comprehensive and accurate assessment. Therefore, by utilizing highly accurate route characteristics and demand data, it is possible to adaptively determine travel plans that meet individual needs, balance diversity and practicality, and ultimately provide personalized target routes.

[0115] Please see Figure 5 This is an optional flowchart illustrating the route generation method provided in an embodiment of this application. Figure 1The step S102 shown can also be implemented by steps S601 to S603, which will be explained in conjunction with the steps:

[0116] S601. Process the entity corresponding to each route using a preset conversion model, and determine the corresponding first preset format feature based on the corresponding route features. Then, process the demand data using the preset conversion model to determine the corresponding second preset format feature.

[0117] In this embodiment, the route generation device can obtain the current user's travel needs data through a human-computer interaction device. The route generation device can use a pre-trained preset conversion model to process the entities and corresponding route features corresponding to each route to obtain each first preset format feature, and use the preset conversion model to process the needs data to obtain a second preset format feature.

[0118] In this embodiment, the route generation device can utilize a large-scale pre-trained model designed for Chinese, capable of generating high-quality text embeddings (dmeta-embedding-zh), parsing user demand data, and forming a clear query vector (second preset format feature). For example, based on the demand data, a user vector q = Encoder(q) is constructed using the trained text embedding model dmeta-embedding-zh. The large-scale pre-trained dmeta-embedding-zh model, capable of generating high-quality text embeddings, can be used to vectorize travel route-related data (including entities within the route and corresponding route features) to obtain the corresponding first preset format feature.

[0119] The demand data includes at least one of the following: the current user's time preference, interest node preference, and route budget.

[0120] S602. Retrieve the top N first preset format features that have the highest similarity to the second preset format features from among the multiple first preset format features.

[0121] In this embodiment, the route generation device can determine the similarity between the first preset format feature and the second preset format feature, and extract the top N first preset format features with the highest similarity from multiple first preset format features. N is an integer greater than 0.

[0122] In this embodiment, similarity can be calculated using Euclidean distance. In other embodiments, similarity can also be determined using other calculation methods.

[0123] For example, the route generation device can use a cosine similarity algorithm to retrieve the top 10 routes with the highest similarity to the second preset format features from multiple first preset format features. The retrieval result is as follows: {k_1,k_2,…,k_k}="Retrieve"(q,K). Here, k_1 is the first route, and k_k is the kth route.

[0124] S603. Based on the N routes and the demand data corresponding to the first N first preset format features, determine the target route.

[0125] In this embodiment of the application, since each first preset format feature corresponds to a route, the route generation device can input the key entities and demand data corresponding to N routes into the large model and output the corresponding target routes.

[0126] For example, the route generation device takes the demand data and the entity inputs corresponding to the top 10 routes and generates the corresponding target routes based on the second largest model, Retrieval-Augmented Generation (RAG), trained on the Chinese fine-tuned version of LLAMA3. The target routes are as follows: a = Generator(LLM)(q,{k1,k2,...,k k}).

[0127] In this embodiment, a preset conversion model is used to process the entity corresponding to each route, and the corresponding route features are used to determine the corresponding first preset format feature. The preset conversion model is then used to process the demand data to determine the corresponding second preset format feature. Among the multiple first preset format features, the top N first preset format features with the highest similarity to the second preset format feature are retrieved. Based on the N routes corresponding to the top N first preset format features and the demand data, the target route is determined. In this way, by comparing the similarity of multiple first preset format features with the second preset format features of the demand data, N routes that meet individual needs, take into account diversity and practicality can be adaptively determined according to user preferences. Furthermore, a personalized target route can be determined using the N routes and the demand data.

[0128] Please see Figure 6 The following is an optional flowchart illustrating the route generation method provided in this application embodiment, which will be described in conjunction with the steps:

[0129] S11. Acquisition and integration of massive amounts of tourism data.

[0130] In this embodiment of the application, the route generation device acquires massive amounts of basic node data from the network and mines and crawls massive amounts of travelogue data from the network.

[0131] S12. Data preprocessing and purification.

[0132] In this embodiment of the application, the route generation device can convert node basic data and travelogue data into text data, and perform noise removal, text standardization, and language detection and filtering on the text data.

[0133] S13, Deep Semantic Analysis and Entity Extraction.

[0134] In this embodiment of the application, the route generation device can perform semantic analysis on text data using a large model and extract key entities of multiple categories from the text data.

[0135] S14, Route Extraction.

[0136] In this embodiment of the application, the route generation device can extract multiple entities with related relationships from multiple key entities to form a route.

[0137] S15. Route deduplication and optimization.

[0138] In this embodiment of the application, the route generation device can perform route hashing on the extracted multiple routes, delete routes with high similarity or repetition, and perform diversity and quality scoring based on the entities included in each route to determine the route characteristics of each route.

[0139] S16. Determine the target route.

[0140] In this embodiment, the route generation device can use a vector model to process the entities corresponding to the route and the corresponding route features to obtain corresponding first preset format features, which are stored in a vector database. It can also use a vector model to process the demand data obtained from the user to obtain corresponding second preset format features, which are also stored in the vector database. Then, using the second preset format features, it retrieves the top N first preset format features with the highest similarity among multiple first preset format features. Finally, it processes the N routes and demand data corresponding to the top N first preset format features using a large model to obtain the target route, which is then fed back to the user.

[0141] Please see Figure 7 This is a schematic diagram of the route generation device provided in the embodiments of this application.

[0142] This application provides a route generation device 800, including: an extraction unit 801 and a determination unit 802.

[0143] Extraction unit 801 is used to extract data from the acquired tourism data using a large model and determine multiple routes;

[0144] The determining unit 802 is used to determine the target route based on the similarity between the relevant data corresponding to the multiple routes and the demand data corresponding to the current user.

[0145] In this embodiment of the application, the extraction unit 801 in the route generation device 800 is used to extract text data in a uniform format from the acquired tourism data using the large model;

[0146] Extract multiple key entities from the text data and establish the relationships between these key entities.

[0147] Among the multiple key entities with related relationships, multiple routes are extracted; wherein, the relevant data of the routes includes at least one of the following: node-related entities, time-related entities, object-related entities, evaluation-related entities, and route mode entities;

[0148] In this embodiment of the application, the extraction unit 801 in the route generation device 800 is used to determine the route features corresponding to the multiple routes respectively.

[0149] In this embodiment of the application, the tourism data includes: multiple node basic data and multiple user-corresponding travelogue data; the extraction unit 801 in the route generation device 800 is used to use the large model to perform text conversion processing on the acquired multiple node basic data and multiple user-corresponding travelogue data to obtain initial text data.

[0150] The initial text data is standardized and filtered to extract text data in a uniform format.

[0151] In this embodiment of the application, the extraction unit 801 in the route generation device 800 is used to standardize and adjust the text content in the text data using a preset language model.

[0152] In this embodiment of the application, the extraction unit 801 in the route generation device 800 is used to extract multiple key entities from the text data based on a preset entity template; wherein, the preset entity template includes: preset node-related entities, preset time-related entities, preset object-related entities, and preset evaluation entities;

[0153] A correlation analysis is performed on multiple key entities to establish the relationships between them.

[0154] In this embodiment of the application, the extraction unit 801 in the route generation device 800 is used to delete routes with a similarity greater than a preset threshold from a plurality of routes based on the similarity between each of the routes.

[0155] In this embodiment of the application, the extraction unit 801 in the route generation device 800 is used to determine the node distribution characteristics corresponding to each of the multiple routes based on the geographical location characteristics and node type characteristics of the node-related entities represented by each route.

[0156] Based on the node identifier of the node-related entity representation corresponding to each route, and the similarity between the node identifier of the node-related entity representation corresponding to other routes, the unique features corresponding to each route are determined.

[0157] Based on the evaluation information of each user represented by the evaluation-related entity corresponding to each route, the evaluation features corresponding to each route are determined.

[0158] The node distribution characteristics, the uniqueness characteristics, and the evaluation characteristics are weighted and summed to determine each route characteristic.

[0159] In this embodiment of the application, the determining unit 802 in the route generation device 800 is used to process the entity corresponding to each route and the corresponding route features using a preset conversion model to determine the corresponding first preset format feature, and to process the demand data using the preset conversion model to determine the corresponding second preset format feature.

[0160] Retrieve from multiple first preset format features the top N first preset format features that have the highest similarity to the second preset format features; N is an integer greater than 0;

[0161] The target route is determined based on the N routes and the demand data corresponding to the first N first preset format features.

[0162] It should be noted that, in the embodiments of this application, if the above-described item information processing method is implemented as a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the embodiments of this application, or the part that contributes to the related technology, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause an item information processing device (which may be a personal computer, etc.) to execute all or part of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), magnetic disks, or optical disks. Thus, the embodiments of this application are not limited to any specific hardware and software combination.

[0163] Correspondingly, embodiments of this application provide a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the steps in the method on one side of the route generation device.

[0164] It should be noted that the descriptions of the storage medium and device embodiments above are similar to the descriptions of the method embodiments above, and have similar beneficial effects. For technical details not disclosed in the storage medium and device embodiments of this application, please refer to the descriptions of the method embodiments of this application for understanding.

[0165] It should be noted that, Figure 8 A schematic diagram of a hardware entity of an electronic device provided in an embodiment of this application, such as... Figure 8 As shown, this application embodiment provides an electronic device 900, including a memory 902 and a processor 901. The memory 902 stores a computer program that can run on the processor 901. When the processor 901 executes the program, it implements the steps in the above-described method, wherein;

[0166] Processor 901 typically controls the overall operation of electronic device 900.

[0167] The memory 902 is configured to store instructions and applications executable by the processor 901, and can also cache data to be processed or already processed (e.g., image data, audio data, voice communication data and video communication data) in the processor 901 and various modules in the electronic device 900. It can be implemented by flash memory or random access memory (RAM).

[0168] Correspondingly, this application also provides a computer program product, including a computer program that can be executed by the processor 901 of the electronic device 900 to complete the steps in the method of the route generation device 800.

[0169] It should be understood that the phrase "one embodiment" or "an embodiment" throughout the specification means that a specific feature, structure, or characteristic related to the embodiment is included in at least one embodiment of this application. Therefore, "in one embodiment" or "in an embodiment" appearing throughout the specification does not necessarily refer to the same embodiment. Furthermore, these specific features, structures, or characteristics can be combined in any suitable manner in one or more embodiments. It should be understood that in the various embodiments of this application, the sequence numbers of the above-described processes do not imply a sequential order of execution; the execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application. The sequence numbers of the above-described embodiments are merely descriptive and do not represent the superiority or inferiority of the embodiments.

[0170] It should be noted that, in this document, 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. Unless otherwise specified, 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 that element.

[0171] In the several embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. The apparatus embodiments described above are merely illustrative. For example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods, such as: multiple units or components can be combined, or integrated into another system, or some features can be ignored or not executed. In addition, the coupling, direct coupling, or communication connection between the various components shown or discussed can be through some interfaces, and the indirect coupling or communication connection of the apparatus or units can be electrical, mechanical, or other forms.

[0172] The units described above as separate components may or may not be physically separate. The components shown as units may or may not be physical units. They may be located in one place or distributed across multiple network units. Some or all of the units may be selected to achieve the purpose of this embodiment according to actual needs.

[0173] In addition, each functional unit in the various embodiments of this application can be integrated into one processing unit, or each unit can be a separate unit, or two or more units can be integrated into one unit; the integrated unit can be implemented in hardware or in the form of hardware plus software functional units.

[0174] Those skilled in the art will understand that all or part of the steps of the above method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it performs the steps of the above method embodiments. The aforementioned storage medium includes various media that can store program code, such as mobile storage devices, read-only memory (ROM), magnetic disks, or optical disks.

[0175] Alternatively, if the integrated units described above are implemented as software functional modules and sold or used as independent products, they can also be stored in a computer-readable storage medium. Based on this understanding, the technical solutions of the embodiments of this application, or the parts that contribute to related technologies, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as mobile storage devices, ROMs, magnetic disks, or optical disks.

[0176] The above description is merely an embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A route generation method, characterized in that, include: The acquired tourism data is used to extract data using a large model to determine multiple routes; The target route is determined based on the similarity between the relevant data corresponding to the multiple routes and the current user's demand data.

2. The route generation method according to claim 1, characterized in that, The process involves using a large model to extract data from the acquired tourism data and determining multiple routes, including: Using the large model, extract text data in a uniform format from the acquired tourism data; Extract multiple key entities from the text data and establish the relationships between these key entities. Among the multiple key entities with related relationships, multiple routes are extracted; wherein the relevant data of the routes includes at least one of the following: node-related entities, time-related entities, object-related entities, evaluation-related entities, and route mode entities.

3. The route generation method according to claim 2, characterized in that, The relevant data also includes: route features; after extracting data from the acquired tourism data using a large model to determine multiple routes, before determining the target route based on the similarity between the relevant data corresponding to each of the multiple routes and the current user's demand data, the method further includes: Determine the route features corresponding to each of the multiple routes.

4. The route generation method according to claim 2, characterized in that, The tourism data includes: basic data from multiple nodes and travelogue data corresponding to multiple users; the process of extracting uniformly formatted text data from the acquired tourism data using the large model includes: Using the large model, the acquired basic data of multiple nodes and the travelogue data corresponding to multiple users are processed by text conversion to obtain initial text data; The initial text data is standardized and filtered to extract text data in a uniform format.

5. The route generation method according to claim 2, characterized in that, After extracting uniformly formatted text data from the acquired tourism data using the large model, and before extracting multiple key entities from the text data and establishing the relationships between the key entities, the method further includes: The text content in the text data is standardized and adjusted using a preset language model.

6. The route generation method according to claim 2, characterized in that, The step of extracting multiple key entities from the text data and establishing the relationships between the key entities includes: Multiple key entities are extracted from the text data based on a preset entity template; wherein the preset entity template includes: preset node-related entities, preset time-related entities, preset object-related entities, and preset evaluation entities; A correlation analysis is performed on multiple key entities to establish the relationships between them.

7. The route generation method according to claim 3, characterized in that, After extracting multiple routes from the multiple key entities with relevant relationships, and before determining the route features corresponding to each of the multiple routes, the method further includes: Based on the similarity between each of the routes, routes with a similarity greater than a preset threshold are deleted.

8. The route generation method according to claim 3, characterized in that, The determination of route features corresponding to the multiple routes includes: For each of the multiple routes, the node distribution characteristics corresponding to each route are determined based on the geographical location features and node type features of the node-related entities represented by each route. Based on the node identifier of the node-related entity representation corresponding to each route, and the similarity between the node identifier of the node-related entity representation corresponding to other routes, the unique features corresponding to each route are determined. Based on the evaluation information of each user represented by the evaluation-related entity corresponding to each route, the evaluation features corresponding to each route are determined. The node distribution characteristics, the uniqueness characteristics, and the evaluation characteristics are weighted and summed to determine each route characteristic.

9. The route generation method according to claim 3, characterized in that, The step of determining the target route based on the similarity between the relevant data corresponding to the multiple routes and the current user's demand data includes: The entity corresponding to each route is processed using a preset conversion model, and the corresponding route features are used to determine the corresponding first preset format features. The demand data is then processed using the preset conversion model to determine the corresponding second preset format features. Retrieve from multiple first preset format features the top N first preset format features that have the highest similarity to the second preset format features; N is an integer greater than 0; The target route is determined based on the N routes and the demand data corresponding to the first N first preset format features.

10. A route generation device, characterized in that, include: The extraction unit is used to extract data from the acquired tourism data using a large model and determine multiple routes; The determining unit is used to determine the target route based on the similarity between the relevant data corresponding to the multiple routes and the demand data corresponding to the current user.

11. An electronic device, characterized in that, It includes a memory and a processor, the memory storing a computer program that can run on the processor, the processor executing the computer program to implement the steps of the method according to any one of claims 1 to 9.

12. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 9.

13. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 9.

Images