Data big model-based one-button travel recommendation method, apparatus and device, and medium

Abstract

The invention discloses a one-key travel recommendation method, device and equipment based on a data big model and a medium, and the method comprises the steps: obtaining a plurality of pieces of user data from a plurality of data sources in a historical time period; obtaining a tourism keyword set; marking and removing a final keyword in each tourism keyword set, and inputting the plurality of tourism keyword sets into the neural network model to be trained to obtain a predicted tourism site corresponding to each tourism keyword set; adjusting neural network parameters of the to-be-trained neural network model according to the final tourist site and the predicted tourist site of each tourist keyword set; and when a preset training requirement is met, storing the newest neural network parameters, and taking the to-be-trained neural network model corresponding to the newest neural network parameters as a one-button tourist site recommendation neural network model. The invention belongs to the field of tourist site recommendation. According to the method, the neural network and the knowledge graph are combined, and the tourist site of the target user can be predicted.

Description

Technical Field

[0001] The present invention relates to the field of tourist spot recommendation, and in particular to a one-click tourist recommendation method, device, equipment and medium based on a big data model. Background Art

[0002] In recent years, the volume of tourism has continued to rise, and holidays have ushered in a peak in passenger flow. Taking the 2024 National Day holiday as an example, the total cross-regional personnel flow in the whole society reached 2.003 billion person-times, becoming a key economic growth point to boost domestic demand. Current travel software such as Xiaohongshu, Mafengwo and Ctrip have integrated scenic spot recommendations for travelers, which to a certain extent facilitates users' itinerary planning.

[0003] The above applications have problems such as advertising interference, information authenticity to be verified, opaque prices, slow customer service response and insufficient after-sales service. There are many advertisements and promotional content on the platform, which greatly affects the user's reading experience. Some customers have "difficulty in choosing" and find it difficult to determine the travel destination among the numerous information. Therefore, how to provide customers with travel destinations in a simple way in accordance with customer needs is an urgent problem to be solved. Summary of the invention

[0004] The present invention solves the technical problem of redundant information recommendation of tourist attractions on various platforms in the prior art by providing a one-click tourist recommendation method, device, equipment and medium based on a big data model, and achieves the technical effect of simply providing tourist attractions to customers.

[0005] In a first aspect, the present invention provides a one-click travel recommendation method based on a big data model, the method comprising: During a historical period, a number of user data are obtained from a number of data sources; Based on the preset tourism knowledge graph, user data is filtered and eliminated to obtain a tourism keyword set, which contains several keywords and corresponds to users one by one; Mark and remove the final keyword in each tourism keyword set, wherein the final keyword is the final tourism location; Input several tourism keyword sets into the neural network model to be trained to obtain the predicted tourism locations corresponding to each tourism keyword set; According to the final tourist destination and the predicted tourist destination of each tourist keyword set, the neural network parameters of the neural network model to be trained are adjusted; When the preset training requirements are met, the latest neural network parameters are saved, and the neural network model to be trained corresponding to the latest neural network parameters is used as the one-click tourist destination recommendation neural network model.

[0006] Further, the method comprises: Obtain the user data of the target user within the target historical time period; Based on the preset tourism knowledge graph, screen and eliminate the user data of the target user to obtain the target tourism keyword set; Input the target tourism keyword set into the one - key tourism location recommendation neural network model to obtain the recommended tourism locations for the target user, and feedback the recommended tourism locations to the client of the target user.

[0007] Furthermore, the loss function of the neural network model to be trained includes: , Among them, is the loss function of the neural network model to be trained, and are both preset weights, is the number of tourism keyword sets, is the th tourism keyword set, is the th tourism keyword set's th real label of the tourism location, is the th tourism keyword set's th predicted probability of the tourism location, is the th tourism location, is the total number of tourism locations, is the base of the natural logarithm, is the decay rate, is the time interval between the time point of the last information interaction of the user corresponding to the th tourism keyword set within the historical time period and the last time point of the historical time period.

[0008] Furthermore, determining the preset tourism knowledge graph includes: Directly obtain the final tourism location corresponding to each user data; According to several final tourism locations, determine the relevance score of each knowledge graph in the knowledge graph database; Take the knowledge graph with the highest relevance score as the preset tourism knowledge graph.

[0009] Furthermore, determining the relevance score of the knowledge graph includes: , Among them, is the relevance score of the th knowledge graph in the knowledge graph database, is the The number of phrases corresponding to the final travel destination in a knowledge graph is the total number of phrases in the nth knowledge graph

[0010] Further, the neural network parameters include: weight matrix, bias vector, parameters of the activation function, learning rate, and decay rate

[0011] Further, the user data is screened and filtered based on a preset travel knowledge graph to obtain a travel keyword set, including: If a phrase in the user data can match any phrase in the preset travel knowledge graph, then the phrase is a keyword; if not, the phrase in the user data is filtered out

[0012] In a second aspect, the present invention provides a one - key travel recommendation device based on a data large model, and the device includes: An acquisition module, configured to obtain a plurality of user data from a plurality of data sources within a historical time period A screening module, configured to screen and filter the user data based on a preset travel knowledge graph to obtain a travel keyword set, where the travel keyword set contains a plurality of keywords, and the travel keyword set corresponds to each user A marking module, configured to mark and filter out the final keyword in each travel keyword set, where the final keyword is the final travel destination A prediction module, configured to input a plurality of travel keyword sets into a neural network model to be trained, and obtain a predicted travel destination corresponding to each travel keyword set An adjustment module, configured to adjust the neural network parameters of the neural network model to be trained according to the final travel destination and the predicted travel destination of each travel keyword set A generation module, configured to save the latest neural network parameters when meeting the preset training requirements, and use the neural network model to be trained corresponding to the latest neural network parameters as a one - key travel destination recommendation neural network model

[0013] In a third aspect, the present invention provides an electronic device, including: A processor A memory for storing instructions executable by the processor wherein the processor is configured to execute to implement the one - key travel recommendation method based on a data large model provided in the first aspect

[0014] Fourthly, the present invention provides a non-transitory computer-readable storage medium, characterized in that when the instructions in the storage medium are executed by a processor of an electronic device, the electronic device can execute the one-key travel recommendation method based on a large data model provided in the first aspect.

[0015] One or more technical solutions provided in the present invention have at least the following technical effects or advantages: The present invention combines a neural network with a knowledge graph to perform one-key prediction of a user's travel destination, avoiding the anxiety caused by the user's difficulty in choosing, avoiding the interference of redundant information, and improving the user experience.

[0016] When designing the loss function of the neural network, the present invention takes into account the influence of data retention time on the user's travel interest, making the model training more efficient and enabling more accurate prediction. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0018] Figure 1 It is a schematic flow chart of the one-key travel recommendation method based on a large data model provided by the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0019] By providing a schematic flow chart of the one-key travel recommendation method in the embodiments of the present invention, the technical problem of redundant information recommendation for travel destinations on various platforms in the prior art is solved.

[0020] The technical solution of the present invention to solve the above technical problem is generally as follows: One - key tourism recommendation method based on a large - data model. The method includes: within a historical time period, obtaining a number of user data from a number of data sources; screening and eliminating the user data based on a preset tourism knowledge graph to obtain a set of tourism keywords, where the set of tourism keywords contains a number of keywords, and the set of tourism keywords corresponds to each user one - to - one; marking and eliminating the final keywords in each set of tourism keywords, where the final keyword is the final tourism destination; inputting a number of sets of tourism keywords into a neural network model to be trained to obtain the predicted tourism destination corresponding to each set of tourism keywords; adjusting the neural network parameters of the neural network model to be trained according to the final tourism destination and the predicted tourism destination of each set of tourism keywords; when the preset training requirements are met, saving the latest neural network parameters, and using the neural network model to be trained corresponding to the latest neural network parameters as a one - key tourism destination recommendation neural network model.

[0021] To better understand the above - mentioned technical solution, the above - mentioned technical solution will be described in detail below in conjunction with the accompanying drawings of the specification and specific implementation manners.

[0022] First, it should be noted that the term "and / or" appearing in this article is merely a description of the association relationship between associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist simultaneously, and B exists alone. In addition, the character " / " in this article generally represents an "or" relationship between the preceding and following associated objects.

[0023] In recent years, the volume of tourism trips has been continuously rising, and the holiday season especially welcomes a peak in the flow of people. Taking the National Day holiday in 2024 as an example, the cumulative cross - regional movement of people in the whole society reached 2.003 billion person - times, becoming a key economic growth point for driving domestic demand. Current tourism travel software such as Xiaohongshu, Mafengwo, and Ctrip integrate scenic spot recommendations for travelers, which to a certain extent facilitates users' itinerary planning. However, there are also problems such as advertising interference, the authenticity of information to be verified, price opacity, slow customer service response, and insufficient after - sales service. Moreover, there are many advertisements and promotional content on the platform, which greatly affects the user's reading experience. And some customers have "choice difficulties" and it is difficult to determine tourist destinations among numerous pieces of information. Therefore, the purpose of the present invention is to provide a one - key recommendation of tourist destinations for customers to avoid choice anxiety.

[0024] The present invention provides a Figure 1 one - key tourism recommendation method based on a large - data model as shown, including steps S11 - S16: Step S11, within a historical time period, obtaining a number of user data from a number of data sources.

[0025] A historical time period refers to a specific time range in the past, which can be several days, weeks, or months. Data sources can include website logs, mobile apps, social media platforms, reservation systems, user feedback channels, third-party APIs, etc.

[0026] User data of each user can be obtained from several of the above data sources. Here, user data refers to personal information or behavioral records, specifically, it can refer to behavioral data, demographic characteristics, preference information, geographical location information, etc.

[0027] Step S12: Based on a preset travel knowledge graph, screen and eliminate user data to obtain a travel keyword set. The travel keyword set contains several keywords, and the travel keyword set corresponds to each user one by one. Specifically, it includes: If a phrase in the user data can match any phrase in the preset travel knowledge graph, then this phrase is a keyword; if not, then eliminate this phrase in the user data.

[0028] It can be understood that there are several phrases in the user data. These phrases can be received as feedback or input by the user themselves. For example, user A asks a certain search website: "Which sea is the most beautiful?"; the search website replies: "Santorini". The question asked by user A to the search website and the reply from the search website are both user data. The phrases to be eliminated can be "sea" and "Santorini" (only 2 examples, and in the above question and answer, there are not only these two phrases).

[0029] Matching can refer to whether the phrase in the user data is the same as the phrase existing in the travel knowledge graph. For example, if a certain phrase of user A is "Santorini" and there is also a phrase "Santorini" in the travel knowledge graph, then the two can be matched.

[0030] The travel knowledge graph can include: geographical location information, tourist attractions, travel services, activities and events, traveler reviews, travel plans, culture and history, etc.

[0031] The above content can be split to obtain several phrases in the travel knowledge graph, and then be matched with the user's phrases.

[0032] Determining the preset travel knowledge graph includes: directly obtaining the final travel location corresponding to each user data; determining the relevance score of each knowledge graph in the knowledge graph database according to several final travel locations; taking the knowledge graph with the highest relevance score as the preset travel knowledge graph.

[0033] Furthermore, it can be set that the users corresponding to the user data circumscribed and obtained in step S11 all went on trips within a preset time period, so as to reduce the interference of redundant data. The preset time period can be a period of time after the historical time period of step S11, and the preset time period is also a past period of time.

[0034] Regarding the travel locations of each user within the preset time period as the final travel locations, it can be understood that the final travel location is a phrase and also a keyword. In the invention, each user's corresponding final travel location is regarded as one. If a user travels to multiple locations within the preset time period, the first travel location is regarded as the final travel location.

[0035] The knowledge graph database includes several types of knowledge graphs, and the knowledge graph with the highest relevance can be used as the preset travel knowledge graph.

[0036] Determining the relevance score of the knowledge graph includes: , where is the relevance score of the th knowledge graph in the knowledge graph database, is the nd number of phrases in the th knowledge graph that contain the phrases corresponding to the final travel location, is the total number of phrases in the

[0037] Step S13, marking and removing the final keywords in each travel keyword set, where the final keyword is the final travel location.

[0038] It can be understood that the final keyword is the real travel result.

[0039] Step S14, inputting several travel keyword sets into the neural network model to be trained, and obtaining the predicted travel location corresponding to each travel keyword set.

[0040] Inputting several travel keyword sets into the neural network model to be trained, each travel keyword set corresponds to a predicted travel location. Matching the predicted travel location of the user with the final travel location to determine whether the predicted travel location is correct.

[0041] Step S15, adjusting the neural network parameters of the neural network model to be trained according to the final travel location and the predicted travel location of each travel keyword set.

[0042] According to whether the predicted travel location is correct and based on the loss function, adjusting and training the neural network model to be trained so that the neural network model to be trained can approach the real travel result as much as possible.

[0043] The loss function of the neural network model to be trained includes: , where is the loss function of the neural network model to be trained, and are both preset weights, is the number of travel keyword sets, is the th travel keyword set, is the th true label of the th travel location in the th travel keyword set, is the predicted probability of the th travel location in the th travel keyword set, is the th travel location, is the base of the natural logarithm, is the decay rate, is the time interval between the time point of the last information interaction of the user corresponding to the th travel keyword set within the historical time period and the last time point of the historical time period.

[0044] The true label refers to the true category or target value of a given data point, that is, the correct answer. For the present invention, it refers to the travel location actually selected by the user. The time point of the last information interaction within the historical time period refers to the time node of the last information record of the user within the historical time period, and the last time point of the historical time period is the end time of the historical time period.

[0045] When designing the loss function, the present invention also considers the influence of the time period on the user's interest degree, improves the prediction accuracy, not only considers the influence of time factors on the user's preference, but also ensures the classification accuracy through the cross-entropy loss, and at the same time realizes the flexible control of the importance of new and old data by using the global adjustment factor.

[0046] Specifically, the longer the time interval between the time point of the last information interaction of the user within the historical time period and the last time point of the historical time period, the smaller the influence of the information obtained from the last information interaction on the selection of the travel location, and vice versa.

[0047] The neural network parameters include: weight matrix, bias vector, parameters of the activation function, learning rate, and decay rate.

[0048] Step S16, when the preset training requirements are met, save the latest neural network parameters, and use the neural network model to be trained corresponding to the latest neural network parameters as the one - key tourism destination recommendation neural network model.

[0049] The preset training requirements can be the number of training times, the matching compliance rate, etc., which will not be elaborated here. When the preset training requirements are met, the one - key tourism destination recommendation neural network model can be used to predict the tourism destinations of users.

[0050] Predicting the tourism destination of the target user based on the one - key tourism destination recommendation neural network model includes: Obtain the user data of the target user in the target historical period; screen and eliminate the user data of the target user based on the preset tourism knowledge graph to obtain the target tourism keyword set; input the target tourism keyword set into the one - key tourism destination recommendation neural network model to obtain the recommended tourism destinations of the target user, and feedback the recommended tourism destinations to the client of the target user. Among them, the target historical period can be the recent few days of the target user.

[0051] In summary, the present invention provides a one - key tourism recommendation method based on a large - data model. The method includes: in a historical period, obtain a number of user data from several data sources; screen and eliminate the user data based on the preset tourism knowledge graph to obtain a tourism keyword set, where the tourism keyword set contains several keywords and corresponds to each user one by one; mark and eliminate the final keyword in each tourism keyword set, where the final keyword is the final tourism destination; input the several tourism keyword sets into the neural network model to be trained to obtain the predicted tourism destinations corresponding to each tourism keyword set; adjust the neural network parameters of the neural network model to be trained according to the final tourism destination and the predicted tourism destination of each tourism keyword set; when the preset training requirements are met, save the latest neural network parameters, and use the neural network model to be trained corresponding to the latest neural network parameters as the one - key tourism destination recommendation neural network model. The present invention combines the neural network with the knowledge graph to perform one - key prediction of the tourism destinations of users, avoiding the anxiety caused by user selection difficulties, avoiding the interference of redundant information, and improving the user experience. When designing the loss function of the neural network in the present invention, the influence of data retention time on users' tourism interests is considered, making the model training more efficient and enabling more accurate prediction.

[0052] Based on the same inventive concept, the present invention provides a one - key tourism recommendation device based on a large - data model. The device includes: An acquisition module, configured to obtain a number of user data from several data sources in a historical period; A screening module for screening and eliminating user data based on a preset tourism knowledge graph to obtain a tourism keyword set. The tourism keyword set contains several keywords and corresponds to each user one by one; A marking module for marking and eliminating the final keywords in each tourism keyword set, where the final keywords are the final tourism destinations; A prediction module for inputting several tourism keyword sets into a neural network model to be trained to obtain the predicted tourism destinations corresponding to each tourism keyword set; An adjustment module for adjusting the neural network parameters of the neural network model to be trained according to the final tourism destinations and the predicted tourism destinations of each tourism keyword set; A generation module for saving the latest neural network parameters when the preset training requirements are met, and using the neural network model to be trained corresponding to the latest neural network parameters as a one-key tourism destination recommendation neural network model.

[0053] Based on the same inventive concept, the present invention also provides an electronic device shown, including: A processor; A memory for storing instructions executable by the processor; Wherein, the processor is configured to execute to implement the one-key tourism recommendation method based on the data large model as provided above.

[0054] Based on the same inventive concept, the present invention also provides a non-transitory computer-readable storage medium. When the instructions in the storage medium are executed by the processor of the electronic device, the electronic device can execute to implement the one-key tourism recommendation method based on the data large model as provided above.

[0055] Since the electronic device introduced in this embodiment is the electronic device used to implement the information processing method in the embodiment of the present invention, based on the information processing method introduced in the embodiment of the present invention, those skilled in the art can understand the specific implementation manners and various variations of the electronic device in this embodiment. Therefore, the specific implementation of how this electronic device implements the method in the embodiment of the present invention will not be described in detail here. As long as the electronic device used by those skilled in the art to implement the information processing method in the embodiment of the present invention belongs to the scope of protection of the present invention.

[0056] Those skilled in the art will appreciate that the embodiments of the present invention can be provided as a method, a system, or a computer program product. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

[0057] The present invention is described with reference to the flowcharts and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the present invention. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, as well as combinations of flows and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, such that the instructions executed by the processor of the computer or other programmable data processing device generate means for implementing the functions specified in Figure 1 one or more of the flows Figure 1 or blocks or combinations of blocks.

[0058] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to operate in a specific manner, such that the instructions stored in the computer-readable memory produce a manufactured article including instruction means that implement the functions specified in Figure 1 one or more of the flows Figure 1 or blocks or combinations of blocks.

[0059] These computer program instructions can also be loaded onto a computer or other programmable data processing device, such that a series of operation steps are performed on the computer or other programmable device to produce a computer-implemented process, and thus the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in Figure 1 one or more of the flows Figure 1 or blocks or combinations of blocks.

[0060] Although the preferred embodiments of the present invention have been described, those skilled in the art can make additional changes and modifications once they learn the basic creative concepts. Therefore, the appended claims are intended to be construed to include the preferred embodiments as well as all changes and modifications that fall within the scope of the present invention.

[0061] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalent technologies, the present invention is also intended to include these modifications and variations.

Claims

1. One-key travel recommendation method based on a large data model, characterized in that, The method includes: Obtaining a number of user data from a number of data sources within a historical time period; Screening and eliminating the user data based on a preset tourism knowledge graph to obtain a tourism keyword set, where the tourism keyword set contains a number of keywords, and the tourism keyword set corresponds to each user one by one; Marking and eliminating the final keywords in each tourism keyword set, where the final keyword is the final tourism destination; Inputting a number of tourism keyword sets into the neural network model to be trained to obtain the predicted tourism destinations corresponding to each tourism keyword set; Adjusting the neural network parameters of the neural network model to be trained according to the final tourism destination and the predicted tourism destination of each tourism keyword set; When the preset training requirements are met, saving the latest neural network parameters, and using the neural network model to be trained corresponding to the latest neural network parameters as a one-key tourism destination recommendation neural network model.

2. The one - key travel recommendation method based on a large - data model according to claim 1, wherein, The method includes: Obtaining the user data of the target user within the target historical time period; Screening and eliminating the user data of the target user based on the preset tourism knowledge graph to obtain a target tourism keyword set; Inputting the target tourism keyword set into the one-key tourism destination recommendation neural network model to obtain the recommended tourism destination of the target user, and feedbacking the recommended tourism destination to the client of the target user.

3. The one - key travel recommendation method based on a large - data model according to claim 1, wherein, The loss function of the neural network model to be trained includes: ， Among them, is the loss function of the neural network model to be trained, and are both preset weights, is the number of travel keyword sets, is the th travel keyword set, is the th real label of the th travel location in the th travel keyword set, is the th predicted probability of the th travel location in the th travel keyword set, is the th travel location, is the total number of travel locations, is the base of the natural logarithm, is the decay rate, is the time interval between the time point of the last information interaction of the user corresponding to the th travel keyword set within the historical time period and the last time point of the historical time period.

4. The one - key travel recommendation method based on a large - data model according to claim 1, wherein, Determining the preset tourism knowledge graph includes: Directly obtaining the final tourism destination corresponding to each user data; Determining the relevance score of each knowledge graph in the knowledge graph database according to a number of final tourism destinations; Using the knowledge graph with the highest relevance score as the preset tourism knowledge graph.

5. The one - key travel recommendation method based on a large - data model according to claim 1, wherein, Determining the relevance score of the knowledge graph includes: ， Among them, is the relevance score of the th knowledge graph in the knowledge graph database, is the number of phrases containing the phrase corresponding to the final travel destination in the th knowledge graph, is the th total number of phrases in the knowledge graph.

6. The one-key travel recommendation method based on a large data model according to claim 1, characterized in that The neural network parameters include: Weight matrix, bias vector, parameters of the activation function, learning rate, and decay rate.

7. The one-key travel recommendation method based on a large data model according to claim 1, wherein Screening and eliminating the user data based on the preset tourism knowledge graph to obtain a tourism keyword set, including: If the phrase in the user data can match any phrase in the preset tourism knowledge graph, then the phrase is a keyword; if not, the phrase in the user data is eliminated.

8. One-key travel recommendation device based on a large data model, characterized in that, The device includes: An acquisition module for obtaining a number of user data from a number of data sources within a historical time period; A screening module for screening and eliminating the user data based on a preset tourism knowledge graph to obtain a tourism keyword set, where the tourism keyword set contains a number of keywords, and the tourism keyword set corresponds to each user one by one; A marking module for marking and eliminating the final keywords in each tourism keyword set, where the final keyword is the final tourism destination; A prediction module for inputting a number of tourism keyword sets into the neural network model to be trained to obtain the predicted tourism destinations corresponding to each tourism keyword set; An adjustment module for adjusting the neural network parameters of the neural network model to be trained according to the final tourism destination and the predicted tourism destination of each tourism keyword set; A generation module, configured to save the latest neural network parameters when the preset training requirements are met, and use the neural network model to be trained corresponding to the latest neural network parameters as a one-key tourist destination recommendation neural network model.

9. An electronic device, characterized in that, Comprising: A processor; A memory for storing executable instructions of the processor; Wherein, the processor is configured to execute to implement the one-key tourism recommendation method based on a large data model according to any one of claims 1 to 7.

10. A non - transitory computer - readable storage medium, characterized in that, When the instructions in the storage medium are executed by the processor of the electronic device, the electronic device is enabled to execute and implement the one-key tourism recommendation method based on a large data model according to any one of claims 1 to 7.

Images