Hotspot-based tourism product matching method, system, device and storage medium

Abstract

The present invention provides a method, system, device, and storage medium for matching tourism products based on hot spots. The method includes: obtaining news information that matches a tourism theme from a set of news information within a preset time period through a classification model; obtaining a popularity parameter of the tourism theme based at least on the number of news information belonging to the same tourism theme and at least one interaction parameter of the news information; pushing a media file package containing at least one news information and relevant tourism product information to a corresponding user based at least on the tourism theme; and pushing relevant tourism product information to the user based at least on a status parameter of the user reading the news information in the media file package. The present invention can tag and store news on and off the website, making it convenient for OTA platforms to use it for promotion scenarios on demand, facilitating attention to relatively real-time tourism hot spot dynamics, timely push of cultural and tourism activities, and increasing the purchase rate of tourism products.

Description

Technical Field

[0001] The present invention relates to the field of online tourism, and in particular to a method, system, device and storage medium for hotspot-based tourism product matching. Background Art

[0002] Online travel agency (OTA) is a specialized term in the travel e-commerce industry. It refers to "travel consumers booking travel products or services online from travel service providers and paying online or offline, enabling travel agents to market or sell products online." The emergence of OTAs has shifted the traditional travel agency sales model onto online platforms, enabling more widespread dissemination of itinerary information and enabling interactive communication that facilitates customer inquiries and bookings.

[0003] Currently, major portals, information platforms, e-commerce apps, and content apps all prominently display popular content or products in prominent locations or sections to capture users' attention. Quickly understanding what the public is searching for, viewing, and following is a key area that many content platforms need to explore. The ability of information flow applications to quickly identify and direct users to these hot topics is a crucial factor influencing user experience. The ability of content platforms to guide users in amplifying and disseminating high-quality hot topics, as well as innovating and recreating them, is key to expanding content marketing and expanding the content ecosystem.

[0004] Hotspot mining is a key technology used by online travel agencies (OTAs) to showcase popular travel destinations such as dining, entertainment, and accommodations. OTAs face significant challenges in quickly and accurately capturing user-uploaded content, analyzing it in a structured manner, aggregating topics and ranking them by popularity, conducting content review, and storing this structured information.

[0005] Moreover, the timing of current product push notifications is relatively random. Product information is only pushed when the user opens the relevant APP, or product information is pushed periodically after collecting and labeling the user's historical data. However, when all platforms use the same form of push notifications, the user's impulse to buy products is paralyzed by the repeated impact of a large number of advertisements, and the user's purchase rate based on traditional information push notifications is gradually declining.

[0006] Therefore, the present invention provides a hotspot-based tourism product matching method, system, device and storage medium. Summary of the Invention

[0007] In response to the problems in the existing technology, the purpose of the present invention is to provide a hotspot-based tourism product matching method, system, equipment and storage medium, which overcomes the difficulties of the existing technology and can label and store news inside and outside the site, making it convenient for OTA platforms to use it for promotion scenarios on demand, facilitating attention to more real-time tourism hotspot dynamics, timely push cultural and tourism activities, and increase the purchase rate of tourism products.

[0008] An embodiment of the present invention provides a hotspot-based tourism product matching method, comprising the following steps:

[0009] Obtaining news information that matches the tourism theme from a collection of news information within a preset time period through a classification model;

[0010] Obtaining a popularity parameter of the tourism theme based at least on the number of news information belonging to the same tourism theme and at least one interaction parameter of the news information;

[0011] Pushing a media file package containing at least one news information and relevant tourism product information to a corresponding user at least according to the tourism theme; and

[0012] At least based on the status parameter of the user reading the news information in the media file package, relevant travel product information is pushed to the user.

[0013] Preferably, the step of obtaining news information that matches the tourism theme from a news information set within a preset time period by using a classification model includes:

[0014] Obtain news information with tourism tags based on the first classification model;

[0015] Classifying news information into tourism topics based on the second classification model; and

[0016] Based on the third classification model, cluster analysis is performed on news under the same tourism theme to obtain different topic clusters under each theme.

[0017] Preferably, the obtaining of news information with a travel tag based on the first classification model includes:

[0018] Use part of the data with certain travel labels as training corpus to train the initial model;

[0019] Use text retrieval models to extract semantic features from massive amounts of news and article data from various sources to create a retrieval database;

[0020] Extract the semantic features of labeled bad cases as query features, and calculate the similarity between the query features and each feature in the retrieval database;

[0021] Set a similarity threshold, select the search titles with high similarity, automatically add the labels annotated by the bad cases as pseudo labels, mix them with the original training set with confirmed travel labels, and retrain the model. The loss function of the pseudo label construction model is as follows:

[0022]

[0023] Where n is the batch size of the data, C is the preset number of categories, f is the maximum value of the network's prediction for labeled data, f' is the maximum value of the network's prediction for unlabeled data, y is the pseudo-label predicted from the labeled data, y' is the pseudo-label predicted from the unlabeled data, and α(t) is a preset variable.

[0024] Preferably, the obtaining of the popularity parameter of the tourism theme based on at least the number of news information belonging to the same tourism theme and at least one interaction parameter of the news information includes:

[0025] Creating a subset of news information belonging to the same tourism theme;

[0026] Obtaining an interaction parameter set for each piece of news information in the news information subset, the interaction parameter set including at least the number of times the news information has been read, forwarded, liked, or collected; and

[0027] The popularity parameter of the news information subset is obtained based on the interaction parameter set according to a weighted algorithm.

[0028] Preferably, the media file package containing at least one piece of news information and relevant tourism product information pushed to the corresponding user at least according to the tourism theme includes:

[0029] Obtaining a travel tag of the user based on the user's historical reading data and / or historical travel data;

[0030] Obtaining N news items with the highest popularity parameters under the travel tag according to the user's travel tag;

[0031] Performing semantic analysis on the news information to obtain tourism keywords, wherein the tourism keywords include tourist attractions, project categories, or tourist seasons;

[0032] Matching at least one travel product according to the travel tag and the travel keyword; and

[0033] A media file package is generated, wherein the media file package includes a plurality of news information and the travel product corresponding to each piece of news information.

[0034] Preferably, the media file package includes a directory showing news information titles, and each of the news information titles corresponds to a text of news information.

[0035] Preferably, the pushing of relevant travel product information to the user based at least on the status parameter of the user reading the news information in the media file package includes:

[0036] Collecting status parameters of the user reading the media file package, wherein the status parameters at least include the reading time and / or interaction parameters of each piece of news information being read by the user;

[0037] Obtaining an interaction parameter of the user to the news information in the media file package based on the reading time and state parameters; and

[0038] When the interaction parameter of the news information meets a preset threshold, the travel product corresponding to the news information is displayed in the media file package and on the display page of the news information.

[0039] Preferably, the pushing of relevant travel product information to the user based at least on the status parameter of the user reading the news information in the media file package includes:

[0040] Collecting status parameters of the user reading the media file package, wherein the status parameters at least include the reading time and / or interaction parameters of each piece of news information being read by the user;

[0041] Obtaining and sorting the user's interaction parameters for the news information in the media file package based on the reading time and status parameters; and

[0042] The travel product corresponding to the news information with the highest interaction parameter is displayed in the media file package and in the directory page of the news information.

[0043] Preferably, the pushing of relevant travel product information to the user based at least on the status parameter of the user reading the news information in the media file package includes:

[0044] Providing a display page to display text of all news information of the media file package;

[0045] Collecting the reading time of each piece of news information in the media file package by the user; and

[0046] When the display page is pulled down to the bottom, the travel products corresponding to the news information with the longest reading time are displayed at the bottom of the display page.

[0047] An embodiment of the present invention further provides a hotspot-based tourism product matching system for implementing the above-mentioned hotspot-based tourism product matching method. The hotspot-based tourism product matching system includes:

[0048] A news classification module, which uses a classification model to obtain news information that matches the tourism theme from a set of news information within a preset time period;

[0049] an interaction parameter module, which obtains a popularity parameter of the tourism theme based on at least the number of news information belonging to the same tourism theme and at least one interaction parameter of the news information;

[0050] A media file module, configured to push a media file package containing at least one news information and relevant tourism product information to a corresponding user at least according to the tourism theme; and

[0051] The tourism product module pushes relevant tourism product information to the user based on at least the status parameter of the user reading the news information in the media file package.

[0052] An embodiment of the present invention further provides a hotspot-based tourism product matching device, comprising:

[0053] processor;

[0054] a memory storing executable instructions for the processor;

[0055] The processor is configured to execute the steps of the above-mentioned hotspot-based tourism product matching method by executing the executable instructions.

[0056] An embodiment of the present invention further provides a computer-readable storage medium for storing a program, which, when executed, implements the steps of the above-mentioned hotspot-based tourism product matching method.

[0057] The purpose of the present invention is to provide a hotspot-based tourism product matching method, system, device and storage medium, which can label and store news inside and outside the site, facilitate OTA platforms to use them for promotion scenarios on demand, facilitate attention to more real-time tourism hotspot dynamics, timely push cultural and tourism activities, and increase the purchase rate of tourism products. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Other features, objects and advantages of the present invention will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings.

[0059] Figure 1 It is a flow chart of the hotspot-based tourism product matching method of the present invention.

[0060] Figures 2 to 5It is a schematic diagram of the implementation process of the hotspot-based tourism product matching method of the present invention.

[0061] Figure 6 It is a module diagram of the hotspot-based tourism product matching system of the present invention.

[0062] Figure 7 It is a structural diagram of the hotspot-based tourism product matching device of the present invention.

[0063] Figure 8 It is a schematic structural diagram of a computer-readable storage medium according to an embodiment of the present invention. DETAILED DESCRIPTION

[0064] The following describes the embodiments of the present application through specific examples. Those skilled in the art can easily understand the other advantages and effects of the present application from the content disclosed in the present application. The present application can also be implemented or applied through different specific embodiments. The details in the present application can also be modified or changed according to different viewpoints and application systems without departing from the spirit of the present application. It should be noted that the embodiments and features in the embodiments of the present application can be combined with each other unless they conflict.

[0065] The following is a detailed description of the embodiments of the present application with reference to the accompanying drawings so that those skilled in the art can easily implement the present application. The present application can be embodied in many different forms and is not limited to the embodiments described herein.

[0066] In the description of this application, reference to the terms "one embodiment," "some embodiments," "examples," "specific examples," or "some examples" means that the specific features, structures, materials, or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of this application. Moreover, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples. In addition, those skilled in the art may combine and integrate different embodiments or examples described in this application, as well as features of different embodiments or examples, unless otherwise contradictory.

[0067] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features being referred to. Thus, a feature identified as "first" or "second" may explicitly or implicitly include at least one such feature. In the context of this application, "plurality" means two or more, unless otherwise specifically defined.

[0068] In order to clearly describe the present application, components not related to the description are omitted, and the same or similar components throughout the specification are given the same reference symbols.

[0069] Throughout this specification, when a device is said to be "connected" to another device, this includes not only "direct connection" but also "indirect connection" with other elements interposed therebetween. Furthermore, when a device is said to "include" a certain component, unless otherwise stated, this does not exclude the inclusion of other components but rather implies that the device may include other components.

[0070] When a device is said to be "on" another device, it may be directly on the other device, but there may also be other devices between it. In contrast, when a device is said to be "directly on" another device, there are no other devices between it.

[0071] Although the terms first, second, etc. are used in some instances herein to represent various elements, these elements should not be limited by these terms. These terms are only used to distinguish one element from another. For example, the first interface and the second interface, etc. are represented. Furthermore, as used in this article, the singular forms "one," "an," and "the" are intended to also include the plural forms, unless there is a contrary indication in the context. It should be further understood that the terms "comprise," "include," and "include" indicate the presence of features, steps, operations, elements, components, items, types, and / or groups, but do not exclude the presence, occurrence, or addition of one or more other features, steps, operations, elements, components, items, types, and / or groups. The terms "or" and "and / or" used herein are interpreted as inclusive, or mean any one or any combination. Therefore, "A, B, or C" or "A, B, and / or C" means "any of the following: A; B; C; A and B; A and C; B and C; A, B, and C." Exceptions to this definition only occur when the combination of elements, functions, steps, or operations is inherently mutually exclusive in some way.

[0072] The technical terms used herein are intended only to refer to specific embodiments and are not intended to limit this application. The singular form used herein also includes the plural form unless the statement explicitly indicates otherwise. The term "comprising" as used in this specification is intended to specify specific features, regions, integers, steps, operations, elements, and / or components and does not exclude the presence or addition of other features, regions, integers, steps, operations, elements, and / or components.

[0073] Although not defined differently, all terms used herein, including technical and scientific terms, have the same meanings as those generally understood by those skilled in the art to which this application belongs. Terms defined in commonly used dictionaries are to be interpreted as having meanings consistent with the relevant technical literature and current teachings, and unless otherwise defined, they should not be overly interpreted as ideal or highly formalized meanings.

[0074] Figure 1 FIG. 1 is a flow chart of the hotspot-based tourism product matching method of the present invention. Figure 1 As shown, an embodiment of the present invention provides a hotspot-based tourism product matching method, comprising the following steps:

[0075] S110 , obtaining news information that meets the tourism theme from a news information set within a preset time period through a classification model.

[0076] S120: Obtain a popularity parameter of the tourism theme based at least on the amount of news information belonging to the same tourism theme and at least one interaction parameter of the news information.

[0077] S130, pushing a media file package containing at least one news information and relevant tourism product information to the corresponding user at least according to the tourism theme.

[0078] S140: Push relevant travel product information to the user based at least on a status parameter of the user reading the news information in the media file package.

[0079] The present invention provides a hotspot-based tourism product matching method. Its improvement lies in utilizing the Transformer technology, widely used in the field of natural language processing (NLP), as a text classification model to obtain a large amount of news related to the tourism field. Then, using text clustering technology, several news items under different themes are obtained. The present invention can aggregate topics with the same attributes under a particular tourism theme, comprehensively evaluate the popularity of tourism news through multiple hot attributes such as topic popularity, topic popularity, and news body hotness, and then rank and display them. This allows for more accurate recognition of information-rich tourism images. Furthermore, the Transformer structure can more easily incorporate multimodal information (such as text, image, and emoticon information) for multimodal learning, resulting in more accurate and richer news tag information. The Transformer model is a seq2seq model proposed by Google Brain in the paper "Attention Is All You Need," published at the end of 2017. The unique feature of the Transformer seq2seq model is its extensive use of a special layer called Self-Attention.

[0080] In a preferred embodiment, step S110 includes:

[0081] S111. Obtain news information with a travel tag based on the first classification model.

[0082] S112. Classify the news information by tourism theme based on the second classification model.

[0083] S113. Based on the third classification model, cluster analysis is performed on news under the same tourism theme to obtain different topic clusters under each theme.

[0084] In a preferred embodiment, step S111 includes:

[0085] Part of the data with certain travel labels is used as training corpus to train the initial model.

[0086] Use text retrieval models to extract semantic features from massive amounts of news and article data from various sources to create a retrieval library.

[0087] The semantic features of labeled bad cases are extracted as query features, and the similarity between the query features and each feature in the retrieval library is calculated.

[0088] Set a similarity threshold, select the search titles with high similarity, automatically add the labels annotated by the bad cases as pseudo labels, mix them with the original training set with confirmed travel labels, and retrain the model. The loss function of the pseudo label construction model is as follows:

[0089]

[0090] Where n is the batch size of the data, C is the preset number of categories, f is the maximum value of the network's prediction for labeled data, f' is the maximum value of the network's prediction for unlabeled data, y is the pseudo-label predicted from the labeled data, y' is the pseudo-label predicted from the unlabeled data, and α(t) is a preset variable.

[0091] In a preferred embodiment, step S120 includes:

[0092] Create a sub-collection of news information belonging to the same tourism theme.

[0093] An interaction parameter set of each news information in the news information subset is obtained, where the interaction parameter set includes at least the number of times the news information is read, forwarded, liked, and collected.

[0094] According to the weighted algorithm and the interaction parameter set, the popularity parameter of the news information subset is obtained.

[0095] In a preferred embodiment, step S130 includes:

[0096] The user's travel tags are obtained according to the user's historical reading data and / or historical travel data.

[0097] According to the user's travel tag, obtain the N news information with the highest popularity parameters under the travel tag.

[0098] Semantic analysis is performed on news information to obtain tourism keywords, which include tourist attractions, project categories or tourist seasons.

[0099] Match at least one travel product based on travel tags and travel keywords.

[0100] A media file package is generated, wherein the media file package includes a plurality of news information and a travel product corresponding to each piece of news information.

[0101] In a preferred embodiment, the media file package includes a directory showing news information titles, each news information title corresponding to a text of news information.

[0102] In a preferred embodiment, step S140 includes:

[0103] The state parameters of the user reading the media file package are collected, and the state parameters at least include the reading time and / or interaction parameters of each news information read by the user.

[0104] The user's interaction parameters for the news information in the media file package are obtained based on the reading time and status parameters.

[0105] When the interaction parameter of the news information meets the preset threshold, the tourism product corresponding to the news information is displayed in the media file package and on the display page of the news information.

[0106] In a preferred embodiment, step S140 includes:

[0107] The state parameters of the user reading the media file package are collected, and the state parameters at least include the reading time and / or interaction parameters of each news information read by the user.

[0108] Based on the reading time and status parameters, the user's interaction parameters for the news information in the media file package are obtained and sorted.

[0109] The tourism products corresponding to the news information with the highest interaction parameter are displayed in the media file package and the directory page of the news information.

[0110] In a preferred embodiment, step S140 includes:

[0111] A display page is provided to display the text of all news information of the media file package.

[0112] Collect the reading time of each news information in the media file package read by the user.

[0113] When the display page is pulled down to the bottom, the travel products corresponding to the news information with the longest reading time are displayed at the bottom of the display page.

[0114] The problem to be solved by the present invention is to use the Transformer technology in deep learning to automatically classify news into its corresponding label system and store structured information of tourism-related news in the database.

[0115] In the actual text recognition process, factors such as the high degree of subjectivity in determining whether a news story is tourism-related, the wide variation in news sources, concise headlines, and limited available information contribute to the difficulty of filtering tourism information. Building a reliable tourism information classification model is a key factor in identifying tourist hotspots. By refining thematic categories and subdividing information sources into categories such as entertainment, automobiles, sports, games, travel, culture, and other, we use an open-source language model to build a multi-classification downstream task, predicting multiple categories. If a category contains a tourism tag greater than a certain threshold, the news is considered tourism-related.

[0116] In a preferred embodiment, when training the downstream classification model, the present invention adds a special sampling strategy to the network. For training samples of different lengths and styles, they are marked according to the length of the text and the source of the text. When using them, weighted sampling is performed according to a certain length threshold, the number of source distributions, etc., which solves the problem of network non-convergence or slow convergence caused by uneven sample distribution.

[0117] In a preferred embodiment, the resulting classification model can better identify tourism-related texts with rich content. In addition, the Transformer structure can more easily integrate multimodal information (such as text images, emojis, video information or voice information of the article) for multimodal learning, thereby obtaining more accurate and richer image label information.

[0118] In a preferred embodiment, after obtaining travel-related news, the news is classified by subject using the BM25 algorithm, and the travel-related news is further classified into finer-grained topics, such as cherry blossoms, skiing, rafting, etc.

[0119] In a preferred embodiment, after obtaining the theme of tourism-related news, the Infomap algorithm is used to perform cluster analysis on the news under the same theme to obtain different topic clusters under the same theme. For example, under the cherry blossom theme, there may be topics about different cherry blossom themed activities held in multiple places, and a representative news headline is selected as the topic title according to certain rules.

[0120] In a preferred embodiment, after obtaining the topic of tourism-related news, the number of news under the comprehensive subject classification, the number of news under the topic classification, and the number of likes, collections, reposts, comments and other related parameters of the news itself are integrated to calculate the comprehensive popularity score of a certain topic according to a certain weight, and sort it according to the popularity score to obtain a list of tourism-related hot spots for the day.

[0121] The present invention uses the learned Transformer model to label and store news inside and outside the site, making it convenient for OTA platforms to use it for promotion scenarios on demand. It can be widely used in tourism recommendations, tourism creation, etc. The present invention can pay attention to relatively real-time tourism hotspot dynamics, and has a positive role in timely pushing cultural and tourism activities and promoting tourism development.

[0122] The present invention can establish a complete rapid iterative model training process:

[0123] Manually annotate some of the data with certain labels as training corpus, train the initial model, predict online data, manually check bad cases (badcases), and add certain labels.

[0124] A text retrieval model is used to extract semantic features from massive amounts of news and article data from various sources to form a retrieval library. The semantic features of labeled badcase are extracted as query features. The similarity between the query feature and each feature in the retrieval library is calculated. A similarity threshold is set, and the retrieval titles with high similarity are taken. The labels annotated with badcase are automatically added as pseudo-labels and mixed with the original training set with certain labels to retrain the model.

[0125] The loss function of the pseudo-label construction model is as follows:

[0126]

[0127] Where n is the batch size, C is the number of predefined categories, f is the maximum value of the network's predictions for labeled data, f' is the maximum value of the network's predictions for unlabeled data, y is the pseudo-label predicted from labeled data, y' is the pseudo-label predicted from unlabeled data, and α(t) is a predefined variable. α(t) determines the role of the cost of unlabeled data in network updates. Choosing an appropriate α(t) is crucial: too large a value degrades performance, while too small a value only marginally improves performance.

[0128] The OTA platform has diverse sources of tourism hotspot information, and the data timeliness is also not synchronized. A data adaptation layer is used to build multi-source asynchronous data processing objects to solve the problem of timed and standardized data output from different access sources in different time periods.

[0129] We build a thematic knowledge base and use specific keywords as triggers to categorize each daily news source into different topics. This generates clusters within each topic, and the size of the clusters is used to measure topic popularity. We also calculate the gradual change in topic popularity over the past two weeks to measure the topic's rising popularity index.

[0130] A language retrieval model is used to extract semantic features from news headlines. Semantic features are then combined with the Levenshtein ratio to evaluate the semantic similarity between headlines. An undirected graph is constructed, with the headlines as nodes and the semantic similarities between them serving as weights between nodes. The Infomap community discovery algorithm is used to identify groups within topic clusters, achieving topic clustering. (Infomap's two-layer encoding approach combines group identification with information encoding. A well-defined grouping results in shorter codes. Therefore, if the code length can be quantified and the grouping that minimizes this length can be found, then a good grouping has been found.)

[0131] The popularity of a topic is measured by the size of the topic cluster. Combined with comprehensive factors such as topic popularity, rising topic popularity, topic popularity, and click popularity of the news itself, we get a comprehensive popularity ranking of the topic. The most complete and beautiful title in the topic is displayed as the topic title.

[0132] The specific implementation of the present invention is as follows:

[0133] Figures 2 to 5 It is a schematic diagram of the implementation process of the hotspot-based tourism product matching method of the present invention.

[0134] Based on the first classification model, news information with tourism tags is obtained. Based on the second classification model, news information is classified into tourism themes. Based on the third classification model, news under the same tourism theme are clustered and analyzed to obtain different topic clusters under each theme.

[0135] Create a subset of news information related to the same tourism theme. Obtain a set of interaction parameters for each piece of news information within the subset. The interaction parameter set includes at least the number of times the news information has been read, forwarded, liked, or saved. Obtain a popularity parameter for the subset of news information based on the interaction parameter set using a weighted algorithm.

[0136] The user's travel tag is obtained based on the user's historical reading data and / or historical travel data. In this embodiment, the user is a gardening enthusiast. Based on the user's travel tag, the N news information with the highest popularity parameters under the travel tag are obtained. Semantic analysis is performed on the news information to obtain tourism keywords, which include tourist attractions, project categories, or tourist seasons. At least one tourism product is matched based on the travel tag and the tourism keywords. A media file package is generated, which includes several news information and tourism products corresponding to each news information. The media file package includes a directory showing the titles of the news information, and each news information title corresponds to the text of a news information. The media file package includes three news information about the Horticultural Expo: "AAAAA", "BBBBB", and "CCCCC". The text of the news information "AAAAA" is "aaaaa..aaaaa", which corresponds to the first tourism product X related to the Horticultural Expo; the text of the news information "BBBBB" is "bbbbb..bbbbb", which corresponds to the second tourism product Y related to the Horticultural Expo; the text of the news information "CCCCC" is "ccccc..ccccc", which corresponds to the third tourism product Z related to the Horticultural Expo.

[0137] In one embodiment, the user's mobile terminal 1 collects the user's reading status parameters of the media file package, and the status parameters include at least the reading time and / or interaction parameters of each news information read by the user. Based on the reading time and status parameters, the user's interaction parameters for the news information in the media file package are obtained. Figure 2 As shown, when the interaction parameter of the news information meets the preset threshold, the tourism products corresponding to the news information are displayed in the media file package and the advertising position 2 in the display page of the news information, so that the most interesting and relevant tourism products can be pushed to the user as soon as the user finishes reading the news information in real time according to the user's interest level in the news information (reading time of a single news information).

[0138] In a preferred embodiment, the user's mobile terminal 1 collects the user's reading status parameters of the media file package, and the status parameters include at least the reading time and / or interaction parameters of each news information read by the user. Based on the reading time and status parameters, the user's interaction parameters for the news information in the media file package are obtained and sorted. Figure 3 As shown, the tourism products corresponding to the news information with the highest interaction parameters are displayed in the media file package and in the advertising position 2 at the bottom of the directory page of the news information, so that the most interesting and relevant tourism products can be pushed to the user as soon as the user finishes reading the news information in real time according to the user's interest level in the news information (reading time of a single news information).

[0139] In a preferred embodiment, a display page is provided by the user's mobile terminal 1 to display the text of all news information in the media file package. The reading time of each news information in the media file package is collected. Figure 4 、 5 As shown, when the display page is pulled down to the bottom by the user, the travel products corresponding to the news information with the longest reading time are displayed in the advertisement position 2 at the bottom of the display page, so that the user can be pushed the most interesting and relevant travel products as soon as he / she finishes reading all the news information in the media file package according to the user's interest in the news information (the reading time of a single news information).

[0140] The present invention can enable users to immediately receive push notifications of travel products related to the news information they are most interested in after just completing targeted news information push, thereby achieving a higher purchase rate of travel products when the user is most excited about the product.

[0141] The hotspot-based tourism product matching method of the present invention can label and store news inside and outside the site, making it convenient for the OTA platform to use it for promotion scenarios on demand, facilitating attention to relatively real-time tourism hotspot dynamics, timely push cultural and tourism activities, and increasing the purchase rate of tourism products.

[0142] Figure 6 Schematic diagram of the module of the hotspot-based tourism product matching system of the present invention. Figure 6 As shown, an embodiment of the present invention further provides a hotspot-based tourism product matching system for implementing the above-mentioned hotspot-based tourism product matching method. The hotspot-based tourism product matching system includes:

[0143] The news classification module 51 obtains news information that meets the tourism theme from a news information set within a preset time period through a classification model.

[0144] The interaction parameter module 52 obtains a popularity parameter of the tourism theme based on at least the number of news information belonging to the same tourism theme and at least one interaction parameter of the news information.

[0145] The media file module 53 pushes a media file package containing at least one news information and relevant tourism product information to the corresponding user at least according to the tourism theme.

[0146] The travel product module 54 pushes relevant travel product information to the user based on at least the status parameter of the user reading the news information in the media file package.

[0147] In a preferred embodiment, the news classification module 51 obtains news information with tourism tags based on the first classification model, classifies the news information into tourism topics based on the second classification model, and performs cluster analysis on news under the same tourism topic based on the third classification model to obtain different topic clusters under each topic.

[0148] In a preferred embodiment, the interaction parameter module 52 establishes a subset of news information related to the same tourism theme. It then obtains an interaction parameter set for each piece of news information within the subset. The interaction parameter set includes at least the number of times the news information has been read, forwarded, liked, or saved. Based on the interaction parameter set, a popularity parameter for the subset of news information is obtained using a weighted algorithm.

[0149] In the first classification model, a portion of data with specific travel tags is used as training material to train the initial model. A text retrieval model is used to extract semantic features from a large amount of news and article data from various sources to form a retrieval database. Semantic features of labeled bad cases are extracted as query features, and the similarity between the query features and each feature in the retrieval database is calculated. A similarity threshold is set, and search titles with high similarity are automatically labeled with the bad case labels as pseudo-labels. These are mixed with the original training set with specific travel tags and the model is retrained. The loss function for the pseudo-label construction model is as follows:

[0150]

[0151] Where n is the batch size of the data, C is the preset number of categories, f is the maximum value of the network's prediction for labeled data, f' is the maximum value of the network's prediction for unlabeled data, y is the pseudo-label predicted from the labeled data, y' is the pseudo-label predicted from the unlabeled data, and α(t) is a preset variable.

[0152] In a preferred embodiment, the media file module 53 obtains a user's travel tags based on the user's historical reading data and / or historical travel data. Based on the user's travel tags, the N news pieces with the highest popularity parameters are obtained. Semantic analysis is performed on the news pieces to obtain travel keywords, which may include tourist attractions, project categories, or travel seasons. At least one travel product is matched based on the travel tags and travel keywords. A media file package is generated, which includes a number of news pieces and a corresponding travel product for each piece of news information.

[0153] In a preferred embodiment, the media file package includes a directory showing news information titles, each news information title corresponding to a text of news information.

[0154] In a preferred embodiment, the tourism product module 54 collects user reading status parameters for the media file package, which include at least the reading time and / or interaction parameters for each piece of news information. Based on the reading time and status parameters, the user's interaction parameters for the news information in the media file package are obtained. When the interaction parameters for the news information meet a preset threshold, the tourism product corresponding to the news information is displayed in the media file package and on the news information display page.

[0155] In a preferred embodiment, the travel product module 54 collects user reading status parameters for the media file package. The status parameters include at least the reading time and / or interaction parameters of each news item. Based on the reading time and status parameters, the user's interaction parameters for the news items in the media file package are obtained and ranked. The travel product corresponding to the news item with the highest interaction parameter is displayed in the media file package and on the news item's catalog page.

[0156] In a preferred embodiment, the travel product module 54 provides a display page showing the text of all news items in the media file package. The user's reading time for each news item in the media file package is collected. When the display page is scrolled down to the bottom, the travel product corresponding to the news item with the longest reading time is displayed at the bottom of the display page.

[0157] The hotspot-based tourism product matching system of the present invention can tag and store news inside and outside the site, making it convenient for the OTA platform to use it for promotion scenarios on demand, facilitating attention to more real-time tourism hotspot dynamics, timely push cultural and tourism activities, and increase the purchase rate of tourism products.

[0158] An embodiment of the present invention further provides a hotspot-based tourism product matching device, comprising a processor and a memory storing executable instructions for the processor. The processor is configured to execute the executable instructions to perform the steps of the hotspot-based tourism product matching method.

[0159] As shown above, the hotspot-based tourism product matching system of this embodiment of the present invention can label and store news inside and outside the site, making it convenient for the OTA platform to use it for promotion scenarios on demand, facilitating attention to more real-time tourism hotspot dynamics, timely push cultural and tourism activities, and increase the purchase rate of tourism products.

[0160] Those skilled in the art will appreciate that various aspects of the present invention may be implemented as systems, methods, or program products. Accordingly, various aspects of the present invention may be implemented in the following forms: entirely in hardware, entirely in software (including firmware, microcode, etc.), or in a combination of hardware and software, collectively referred to herein as "circuits," "modules," or "platforms."

[0161] Figure 7 This is a schematic diagram of the structure of the hotspot-based tourism product matching device of the present invention. Figure 7 An electronic device 600 according to this embodiment of the present invention will be described. Figure 7 The electronic device 600 shown is merely an example and should not limit the functions and scope of use of the embodiments of the present invention.

[0162] like Figure 7 As shown, electronic device 600 is implemented as a general-purpose computing device. Components of electronic device 600 may include, but are not limited to, at least one processing unit 610, at least one storage unit 620, a bus 630 connecting different platform components (including storage unit 620 and processing unit 610), and a display unit 640.

[0163] The storage unit stores program codes, which can be executed by the processing unit 610, so that the processing unit 610 executes the steps according to various exemplary embodiments of the present invention described in the electronic prescription circulation processing method section of this specification. For example, the processing unit 610 can execute the following steps: Figure 1 Follow the steps shown in .

[0164] The storage unit 620 may include a readable medium in the form of a volatile storage unit, such as a random access memory unit (RAM) 6201 and / or a cache memory unit 6202 , and may further include a read-only memory unit (ROM) 6203 .

[0165] The storage unit 620 may also include a program / utility 6204 having a set (at least one) of program modules 6205, such program modules 6205 including but not limited to: an operating system, one or more application programs, other program modules, and program data, each of which or some combination may include an implementation of a network environment.

[0166] Bus 630 may represent one or more of several types of bus structures, including a memory bus or memory controller, a peripheral bus, an accelerated graphics port, a processing unit, or a local bus using any of a variety of bus architectures.

[0167] The electronic device 600 can also communicate with one or more external devices 700 (e.g., a keyboard, a pointing device, a Bluetooth device, etc.), one or more devices that enable a user to interact with the electronic device 600, and / or any device that enables the electronic device 600 to communicate with one or more other computing devices (e.g., a router, a modem, etc.). Such communication can occur via an input / output (I / O) interface 650. Furthermore, the electronic device 600 can communicate with one or more networks (e.g., a local area network (LAN), a wide area network (WAN), and / or a public network such as the Internet) via a network adapter 660. The network adapter 660 can communicate with other modules of the electronic device 600 via the bus 630. It should be understood that, although not shown in the figures, other hardware and / or software modules can be used in conjunction with the electronic device 600, including but not limited to microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage platforms.

[0168] An embodiment of the present invention further provides a computer-readable storage medium for storing a program that, when executed, implements the steps of the hotspot-based travel product matching method. In some possible implementations, various aspects of the present invention may also be implemented in the form of a program product, which includes program code. When the program product is executed on a terminal device, the program code is used to cause the terminal device to execute the steps according to various exemplary embodiments of the present invention described in the aforementioned electronic prescription circulation processing method section of this specification.

[0169] As shown above, the hotspot-based tourism product matching system of this embodiment of the present invention can label and store news inside and outside the site, making it convenient for the OTA platform to use it for promotion scenarios on demand, facilitating attention to more real-time tourism hotspot dynamics, timely push cultural and tourism activities, and increase the purchase rate of tourism products.

[0170] Figure 8 Schematic diagram of the structure of the computer readable storage medium of the present invention. Figure 8 , a program product 800 for implementing the above method according to an embodiment of the present invention is described. The program product 800 may be a portable compact disc read-only memory (CD-ROM) and include program code, and may be run on a terminal device, such as a personal computer. However, the program product of the present invention is not limited thereto. In this document, a readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0171] The program product may employ any combination of one or more readable media. The readable medium may be a readable signal medium or a readable storage medium. The readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or component, or any combination thereof. More specific examples (a non-exhaustive list) of readable storage media include: an electrical connection having one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof.

[0172] Computer-readable storage media may include a data signal propagated in baseband or as part of a carrier wave, which carries readable program code. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. The readable storage medium may also be any readable medium other than a readable storage medium, which may send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, apparatus, or device. The program code contained on the readable storage medium may be transmitted using any suitable medium, including but not limited to wireless, wired, optical cable, RF, etc., or any suitable combination thereof.

[0173] The program code for performing the operations of the present invention may be written in any combination of one or more programming languages, including object-oriented programming languages such as Java, C++, and the like, as well as conventional procedural programming languages such as "C" or similar programming languages. The program code may be executed entirely on the user computing device, partially on the user device, as a stand-alone software package, partially on the user computing device and partially on a remote computing device, or entirely on a remote computing device or server. In the case of a remote computing device, the remote computing device may be connected to the user computing device via any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computing device (e.g., via the Internet using an Internet service provider).

[0174] In summary, the purpose of the present invention is to provide a hotspot-based tourism product matching method, system, device and storage medium, which can label and store news inside and outside the site, facilitate OTA platforms to use them for promotion scenarios on demand, facilitate attention to more real-time tourism hotspot dynamics, timely push cultural and tourism activities, and increase the purchase rate of tourism products.

[0175] The above is a further detailed description of the present invention in conjunction with specific preferred embodiments, and the specific implementation of the present invention should not be considered to be limited to these descriptions. For those skilled in the art of the present invention, without departing from the concept of the present invention, several simple deductions or substitutions can be made, which should be considered to fall within the scope of protection of the present invention.

Claims

1. A tourism product matching method based on hotspots, characterized in that: The following steps are involved: Based on the first classification model, news information with tourism tags is obtained, based on the second classification model, news information is classified into tourism themes, and based on the third classification model, news under the same tourism theme is clustered and analyzed to obtain different topic clusters under each of the themes. The news information with tourism tags obtained based on the first classification model includes: using part of the data with the determined tourism tags as training corpus to train the initial model; using the text retrieval model to extract semantic features from a large amount of news and article data from various sources to make a retrieval library; extracting the semantic features of labeled bad cases as query features, and calculating the similarity between the query features and each feature in the retrieval library; setting a similarity threshold, taking the search titles with high similarity, automatically adding the labels marked by the bad cases as pseudo labels, and mixing them with the original training set with the determined tourism tags, and retraining the model, wherein the loss function of the pseudo label construction model is as follows: Where n is the batch size of the data, C is the preset number of categories, f is the maximum value of the network's prediction for labeled data, f' is the maximum value of the network's prediction for unlabeled data, y is the pseudo-label predicted from the labeled data, y' is the pseudo-label predicted from the unlabeled data, and α(t) is a preset variable. Establish a subset of news information belonging to the same tourism theme; obtain an interaction parameter set for each piece of news information in the news information subset, the interaction parameter set including at least the number of times the news information has been read, forwarded, liked, and collected; and obtain a popularity parameter for the news information subset based on the interaction parameter set according to a weighted algorithm; Pushing a media file package containing at least one news information and relevant tourism product information to a corresponding user at least according to the tourism theme; and The relevant travel product information is pushed to the user based at least on the status parameters of the user reading the news information in the media file package, wherein the status parameters at least include the reading time and / or interaction parameters of the user reading each piece of the news information in the media file package.

2. The hotspot-based tourism product matching method according to claim 1, wherein: The method of pushing a media file package containing at least one piece of news information and relevant tourism product information to a corresponding user at least according to the tourism theme includes: Obtaining a travel tag of the user based on the user's historical reading data and / or historical travel data; Obtaining N news items with the highest popularity parameters under the travel tag according to the user's travel tag; Performing semantic analysis on the news information to obtain tourism keywords, wherein the tourism keywords include tourist attractions, project categories, or tourist seasons; Matching at least one travel product according to the travel tag and the travel keyword; and A media file package is generated, wherein the media file package includes a plurality of news information and the travel product corresponding to each piece of news information.

3. The hotspot-based tourism product matching method according to claim 2, wherein: The media file package includes a directory showing news information titles, and each of the news information titles corresponds to a text of news information.

4. The hotspot-based tourism product matching method according to claim 3, wherein: The pushing of relevant travel product information to the user based at least on the status parameter of the user reading the news information in the media file package includes: Collect status parameters of users reading media file packages; Obtaining an interaction parameter of the user to the news information in the media file package based on the reading time and state parameters; and When the interaction parameter of the news information meets a preset threshold, the travel product corresponding to the news information is displayed in the media file package and on the display page of the news information.

5. The hotspot-based tourism product matching method according to claim 3, wherein: The pushing of relevant travel product information to the user based at least on the status parameter of the user reading the news information in the media file package includes: Collect status parameters of users reading media file packages; Obtaining and sorting the user's interaction parameters for the news information in the media file package based on the reading time and status parameters; and The travel product corresponding to the news information with the highest interaction parameter is displayed in the media file package and in the directory page of the news information.

6. The hotspot-based tourism product matching method according to claim 1, wherein: The pushing of relevant travel product information to the user based at least on the status parameter of the user reading the news information in the media file package includes: Providing a display page to display text of all news information of the media file package; Collecting the reading time of each piece of news information in the media file package by the user; and When the display page is pulled down to the bottom, the travel products corresponding to the news information with the longest reading time are displayed at the bottom of the display page.

7. A hotspot-based tourism product matching system, used to implement the hotspot-based tourism product matching method according to claim 1, characterized in that: include: The news classification module obtains news information with tourism tags based on the first classification model, classifies the news information into tourism themes based on the second classification model, and performs cluster analysis on news under the same tourism theme based on the third classification model to obtain different topic clusters under each of the themes. The news information with tourism tags obtained based on the first classification model includes: using part of the data with specific tourism tags as training corpus to train the initial model; using the text retrieval model to extract semantic features from a large amount of news and article data from various sources to make a retrieval library; extracting the semantic features of labeled bad cases as query features, and calculating the similarity between the query features and each feature in the retrieval library; setting a similarity threshold, taking the search titles with high similarity, automatically adding the labels marked by the bad cases as pseudo labels and mixing them with the original training set with specific tourism tags, and retraining the model. The loss function of the pseudo label construction model is as follows: Where n is the batch size of the data, C is the preset number of categories, f is the maximum value of the network's prediction for labeled data, f' is the maximum value of the network's prediction for unlabeled data, y is the pseudo-label predicted from the labeled data, y' is the pseudo-label predicted from the unlabeled data, and α(t) is a preset variable. An interaction parameter module is configured to establish a subset of news information belonging to the same tourism theme; obtain an interaction parameter set for each piece of news information within the news information subset, the interaction parameter set including at least the number of times the news information has been read, forwarded, liked, and collected; and obtain a popularity parameter for the news information subset based on the interaction parameter set according to a weighted algorithm; a media file module for pushing a media file package containing at least one news information and relevant tourism product information to a corresponding user at least according to the tourism theme; and The tourism product module pushes relevant tourism product information to the user based on at least the status parameters of the user reading the news information in the media file package, and the status parameters at least include the reading time and / or interaction parameters of the user reading each piece of news information in the media file package.

8. A tourism product matching device based on hotspots, characterized in that: include: processor; a memory storing executable instructions for the processor; The processor is configured to execute the steps of the hotspot-based tourism product matching method according to any one of claims 1 to 6 by executing the executable instructions.

9. A computer-readable storage medium for storing a program, characterized in that: When the program is executed by a processor, the steps of the hotspot-based tourism product matching method according to any one of claims 1 to 6 are implemented.

Images