A big data management system based on pet classification tags

Abstract

The application discloses a big data management system based on pet classification labels, and relates to the technical field of big data management.The system comprises a content data collection module, a pet classification label module, a user portrait module, a content recommendation module, a geographic position service module, an interactive statistical module and a data storage module.The pet classification label module identifies pet breeds through image recognition technology, and generates a three-level classification label system comprising pet category labels, pet breed labels and content theme labels.The content recommendation module realizes accurate recommendation by combining pet classification labels and user interest labels.The interactive statistical module calculates a comprehensive heat value by using a weighting method.The application realizes fine classification management and accurate recommendation and distribution of pet content.

Description

Technical Field

[0002] This invention relates to the field of big data management technology, specifically to a big data management system based on pet classification tags. Background Technology

[0004] With the improvement of people's living standards and the popularization of pet ownership, pets have become important members of many families. Pet owners want to record and share their pets' daily lives, and at the same time, they also want to acquire professional pet care knowledge and meet like-minded pet lovers. This has spurred the rapid development of pet-related social platforms.

[0005] Existing pet-related content platforms have the following technical problems in content management:

[0006] First, the content classification management is rudimentary. Existing platforms only classify pet content based on simple tags, lacking a multi-level classification system based on pet breeds and content themes. This results in low content retrieval efficiency, making it difficult for users to quickly find content of interest.

[0007] Second, the accuracy of content recommendations is insufficient. The existing recommendation algorithms have failed to fully take into account the special characteristics of the pet industry and cannot make accurate recommendations based on users' preferred pet breeds and pet-related topics, resulting in a poor user experience.

[0008] Third, the local content distribution capability is weak. Pet owners have a strong need for offline social interaction and hope to meet pet lovers in the same city. However, existing platforms lack effective geolocation services to support them and cannot achieve accurate local content distribution.

[0009] Fourth, the utilization of interactive data is insufficient. Existing platforms lack systematic statistical analysis of users' likes, comments, and collection behavior data, and have failed to transform interactive data into a basis for content quality assessment and recommendation ranking.

[0010] Therefore, a big data management system based on pet classification tags is needed to solve the above-mentioned technical problems. Summary of the Invention

[0012] The purpose of this invention is to provide a big data management system based on pet classification tags to solve the technical problems existing in pet content platforms in terms of content classification, accurate recommendation, local distribution, and interaction statistics.

[0013] To achieve the above objectives, the present invention adopts the following technical solution:

[0014] A big data management system based on pet classification tags includes:

[0015] The content data collection module is used to collect pet content data published by users. Pet content data includes text and image data, video data, and nine-grid image data. Pet content data includes pet image information, text description information, publication time information, and publication location information.

[0016] The pet classification tag module, connected to the content data acquisition module, is used to classify and tag pet content data. It includes a pet breed identification unit and a tag generation unit. The pet breed identification unit uses image recognition technology to identify pet breed information in pet content data. The tag generation unit generates corresponding pet classification tags based on the identification results. Pet classification tags include pet category tags, pet breed tags, and content theme tags.

[0017] The user profile module is used to build user interest tags based on users' browsing behavior data, interaction behavior data, and attention behavior data. User interest tags include preferred pet category tags, preferred pet breed tags, and preferred content theme tags.

[0018] The content recommendation module, connected to the pet category tag module and the user profile module, is used to match content based on pet category tags and user interest tags to generate a personalized recommended content stream.

[0019] The geolocation service module is used to obtain the user's geolocation information and filter pet content data within a specified range based on the geolocation information;

[0020] The interaction statistics module is used to collect interaction data for pet content, including likes, comments, favorites, and feeding data, and to calculate a comprehensive popularity value based on the interaction data.

[0021] The data storage module is used to store pet content data, pet category tags, user interest tags, and interaction data.

[0022] Furthermore, the pet category labels include cats, dogs, and exotic pets, with the exotic pet category including birds and rabbits;

[0023] Content tags include daily record tags, training technique tags, health science tags, and cute pet stories tags.

[0024] Furthermore, the content recommendation module includes a priority recommendation unit, a popularity ranking recommendation unit, and a breed matching recommendation unit. The priority recommendation unit prioritizes displaying pet content data published by authors that the user has followed. The popularity ranking recommendation unit sorts pet content data according to the overall popularity value. The breed matching recommendation unit matches corresponding pet content data according to the preferred pet breed tags in the user's interest tags.

[0025] Furthermore, the comprehensive popularity value is calculated as follows: the number of likes is multiplied by the first weight coefficient, the number of comments is multiplied by the second weight coefficient, the number of favorites is multiplied by the third weight coefficient, and the number of feeds is multiplied by the fourth weight coefficient. The four products are then added together to obtain the comprehensive popularity value, where the first weight coefficient is 1, the second weight coefficient is 2, the third weight coefficient is 3, and the fourth weight coefficient is 5.

[0026] Furthermore, the geolocation service module includes a positioning unit, a distance filtering unit, and a city switching unit. The positioning unit obtains the user's current geographic coordinates, the distance filtering unit filters pet content data within a preset filtering radius, which includes three levels: 5 kilometers, 10 kilometers, and 20 kilometers, and the city switching unit allows users to manually switch target cities.

[0027] Furthermore, it also includes a content review module, which consists of a machine review unit and a human review unit. The machine review unit uses artificial intelligence technology to detect illegal content in pet content data, while the human review unit reviews the review results of the machine review unit. Only pet content data that passes the review can enter the content recommendation module for distribution.

[0028] Furthermore, it also includes a pet profile management module, which records the user's pet information, including the date of first pet ownership, pet breed information, pet age information, and vaccination status information. The pet profile management module supports setting the visibility range of pet information, including public visibility, visible to friends, and visible only to the user.

[0029] Furthermore, it also includes a contribution value calculation module, which calculates user contribution value based on users' note-posting, liking, commenting, and sharing behaviors. User contribution value is used to redeem platform benefits.

[0030] Furthermore, it also includes a Featured Content module, which is used to filter pet content data whose overall popularity value reaches a preset threshold of 100. Pet content data that reaches the preset threshold will be displayed at the top of the Featured section.

[0031] Furthermore, the data storage module adopts a distributed storage architecture, including a content data storage unit, a tag index storage unit, and a user data storage unit. The tag index storage unit uses an inverted index structure to store the mapping relationship between pet classification tags and pet content data.

[0032] The beneficial effects of this invention are as follows:

[0033] First, this invention achieves multi-level classification management of pet content through a pet classification tag module, including three dimensions: pet category tags, pet breed tags, and content theme tags, thereby improving content retrieval efficiency and management precision.

[0034] Secondly, this invention constructs user interest tags through a user profile module and combines them with pet classification tags to achieve accurate content recommendations, thereby improving user experience and content distribution efficiency.

[0035] Third, this invention realizes local content distribution based on user location through a geolocation service module, supports multiple filtering radii and city switching functions, and meets the local social needs of pet owners.

[0036] Fourth, this invention uses an interactive statistics module to collect data on likes, comments, favorites, and donations, and uses a weighted calculation method to obtain a comprehensive popularity value, providing data support for content quality assessment and recommendation ranking.

[0037] Fifth, this invention records users' pet information through a pet profile management module, supports privacy visibility settings, and enhances user stickiness and platform belonging. Attached Figure Description

[0039] Figure 1 This is a schematic diagram of the overall architecture of the big data management system based on pet classification tags of the present invention;

[0040] Figure 2 This is a schematic diagram of the pet classification tag module of the present invention;

[0041] Figure 3 This is a schematic diagram of the structure of the content recommendation module of this invention;

[0042] Figure 4 This is a schematic diagram of the geolocation service module of the present invention;

[0043] Figure 5 This is a schematic diagram of the data processing flow of the present invention;

[0044] Figure 6 This is a schematic diagram of the comprehensive heat value calculation process of the present invention. Detailed Implementation

[0046] The present invention will now be described in further detail with reference to the accompanying drawings and specific embodiments.

[0047] like Figure 1As shown, the present invention provides a big data management system based on pet classification tags. The system includes a content data collection module, a pet classification tag module, a user profile module, a content recommendation module, a geolocation service module, an interaction statistics module, and a data storage module.

[0048] I. Content Data Acquisition Module

[0049] The content data collection module is used to collect pet content data published by users. The types of pet content data include text and image data, video data, and nine-grid image data. Among them, text and image data allows users to upload 1 to 9 images along with text descriptions.

[0050] Video data supports users uploading short videos with a duration of no more than 3 minutes;

[0051] The nine-grid image data is a combination display of nine images arranged in three rows and three columns.

[0052] Pet content data includes the following information:

[0053] (1) Pet image information: including pet photos and pet images in videos uploaded by users;

[0054] (2) Text description information: title text, body description, and topic tags entered by the user;

[0055] (3) Publication time information: The specific timestamp of the content publication;

[0056] (4) Location information: The geographic coordinates of the user when posting content.

[0057] The content data acquisition module also provides basic editing functions, including image cropping, filter addition, text overlay, and sticker addition.

[0058] II. Pet Category Tag Module

[0059] like Figure 2 As shown, the pet classification tag module is connected to the content data acquisition module and is used to classify and tag pet content data. This module includes a pet breed identification unit and a tag generation unit.

[0060] The pet breed identification unit analyzes pet images in pet content data using image recognition technology to identify the pet's breed information. Specifically, the pet image is input into a pre-trained convolutional neural network model, and the model outputs the pet breed identification result.

[0061] The tag generation unit generates corresponding pet classification tags based on the recognition results. The pet classification tags adopt a three-level classification system:

[0062] The first level is the pet category label, which includes cats, dogs, and exotic pets. Exotic pets specifically include birds and rabbits.

[0063] The second level is the pet breed label, which is further subdivided based on the pet category label. For example, for cats, the breed labels include Ragdoll, British Shorthair, American Shorthair, Orange Tabby, and Tabby. For dogs, the breed labels include Golden Retriever, Labrador Retriever, Corgi, Poodle, and Border Collie.

[0064] The third level is content theme tags, which are categorized according to the theme type of the content. These include daily record tags, training technique tags, health science tags, and cute pet stories tags. Daily record tags are used to mark content related to pet eating, playing, and sleeping.

[0065] Training tips tags are used to label teaching content such as potty training and handshake commands for pets;

[0066] Health information labels are used to indicate information about vaccination and dietary precautions;

[0067] The "Cute Pet Stories" tag is used to label funny behaviors and pet interactions.

[0068] III. User Profile Module

[0069] The user profiling module is used to construct user interest tags based on user behavior data. The collected user behavior data includes three categories:

[0070] (1) Browsing behavior data: Records of pet content viewed by users, including content identifiers, browsing duration, and browsing frequency;

[0071] (2) Interactive behavior data: User interaction records with pet content, including likes, comments, favorites, and feeding records;

[0072] (3) Follow behavior data: Records of authors followed by users, including the identifier of the followed author and the time of following.

[0073] Based on the above behavioral data, the user profiling module constructs user interest tags, which include three dimensions:

[0074] (1) Pet category tags: determined based on the distribution of pet categories in the content that users browse and interact with;

[0075] (2) Pet breed preference tags: determined based on the distribution of pet breeds in the content browsed and interacted with by users;

[0076] (3) Preference content topic tags: Determined based on the distribution of topic types in the content that users browse and interact with.

[0077] IV. Content Recommendation Module

[0078] like Figure 3 As shown, the content recommendation module is connected to the pet category tag module and the user profile module, and is used to match content based on pet category tags and user interest tags to generate personalized recommended content streams.

[0079] The content recommendation module includes three recommendation units:

[0080] (1) Priority recommendation unit: The pet content data published by authors that the user has followed will be displayed at the top of the content stream. The specific implementation method is: obtain the user's following list, query the content published by each author in the following list, sort them in reverse order of publication time and output them first.

[0081] (2) Popularity ranking recommendation unit: Pet content data is ranked according to the comprehensive popularity value. The higher the comprehensive popularity value, the higher the ranking of the content in the recommendation list.

[0082] (3) Breed matching recommendation unit: Match corresponding pet content data according to the preferred pet breed tags in the user's interest tags. The specific implementation method is: obtain the user's preferred pet breed tags, retrieve pet content with the same breed tags from the data storage module, and make recommendations.

[0083] The recommendation priority of the content recommendation module is as follows: the results of the "Focus Priority Recommendation" unit have the highest priority, followed by the "Popularity Ranking Recommendation" unit, and the results of the "Product Matching Recommendation" unit are used as supplementary recommendations.

[0084] V. Location Service Module

[0085] like Figure 4 As shown, the geolocation service module is used to obtain the user's geolocation information and filter pet content data within a specified range based on the geolocation information. This module includes a location unit, a distance filtering unit, and a city switching unit.

[0086] The positioning unit obtains the user's current geographic coordinates, including longitude and latitude values, and the positioning method combines GPS positioning and network positioning.

[0087] The distance filtering unit filters pet content data within a preset filtering radius. The filtering radius offers three levels: 5 km, 10 km, and 20 km. The default filtering radius is 5 km. The filtering method is to calculate the spherical distance between the user's current coordinates and the coordinates where each pet content is published, and filter out content whose distance is less than the filtering radius.

[0088] The city switching unit allows users to manually switch target cities. Users can select a target city through the city selection interface, and the system will display pet content data within that city.

[0089] VI. Interactive Statistics Module

[0090] The interaction statistics module is used to collect interaction data for pet content and calculate the overall popularity value based on the interaction data.

[0091] Interactive data includes four categories:

[0092] (1) Likes data: Record the number of likes received for each pet content. Each user can only like the same content once.

[0093] (2) Comment data: Records the number of comments received for each pet content. Comments support emojis and tagging users.

[0094] (3) Collection data: Records the number of times each pet content is collected, and users can classify and manage the collected content;

[0095] (4) Feeding data: Record the number of times virtual props are fed to each pet content. Virtual props include cat food props and dog bone props.

[0096] like Figure 6 As shown, the calculation method for the overall heat value is as follows:

[0097] Let the number of likes be The number of comments is The number of collections is The number of feeders is Overall popularity value .

[0098] The formula for calculating the overall popularity value is:

[0100] in:

[0101] This represents the overall popularity score, which is a positive integer.

[0102] This represents the number of likes, a non-negative integer, indicating the total number of likes received for this content;

[0103] The comment count is a non-negative integer, representing the total number of comments received for this content.

[0104] This represents the number of times the content has been saved or added to favorites; it is a non-negative integer.

[0105] The number of virtual items fed to this content is a non-negative integer, representing the total number of times this content has received virtual item feeds.

[0106] This represents the first weight coefficient, with a value of 1, corresponding to the weight of the "like" action;

[0107] This represents the second weighting coefficient, with a value of 2, corresponding to the weight of the commenting behavior;

[0108] This represents the third weighting coefficient, with a value of 3, corresponding to the weight of the collection behavior;

[0109] This represents the fourth weighting coefficient, with a value of 5, corresponding to the weight of the feeding behavior.

[0110] The principle for setting the weight coefficient is: feeding behavior requires points, which represents the user's high recognition of the content, so it has the highest weight;

[0111] The act of saving indicates that users intend to view the content again later, representing that the content has high value, and its weight is secondary;

[0112] Commenting requires users to invest time in writing text, and therefore carries higher weight than liking.

[0113] Liking is the lowest-cost and lowest-weighted action.

[0114] VII. Content Review Module

[0115] The content moderation module is used to conduct compliance reviews of pet content data posted by users. This module includes machine review units and human review units.

[0116] The machine review unit uses artificial intelligence technology to detect illegal content in pet-related data. The detection scope includes: animal abuse scenes, false medical advertisements, sensitive information, and prohibited language. The machine review unit uses a pre-trained content review model to analyze images and text and output the review results.

[0117] The human review unit reviews the results of the machine review unit. Content that the machine review determines to be suspected of being in violation will enter the human review queue for manual judgment by reviewers.

[0118] Only approved pet content data can be distributed through the content recommendation module. Content that fails the review will be blocked, and the system will send a notification to the user explaining the reason for the violation.

[0119] VIII. Pet Record Management Module

[0120] The pet profile management module is used to record users' pet information, which includes:

[0121] (1) First pet ownership time: Records the time when the user started raising the current pet;

[0122] (2) Pet breed information: Record the breed type of the pet;

[0123] (3) Pet age information: Record the pet's age;

[0124] (4) Vaccination status information: Record the pet's vaccination status.

[0125] The pet profile management module supports setting the visibility of pet information. The visibility range includes three levels: public, friends, and only. Public means that all users can view it.

[0126] "Visible to friends" means that only users who follow each other can view it.

[0127] "Only visible to myself" means that only the user can view it.

[0128] IX. Contribution Value Calculation Module

[0129] The contribution value calculation module calculates a user's contribution value based on their behavior, including posting notes, liking, commenting, and sharing.

[0130] User contribution points can be used to redeem platform benefits, including: pinned notes, gift packs, and priority access to expert Q&A.

[0131] 10. Featured Content Module

[0132] The Featured Content module is used to filter high-quality pet content data. The filtering rule is: pet content data with a comprehensive popularity value that reaches a preset threshold will be included in the Featured section. The preset threshold is 100.

[0133] Content featured in the Featured section will be displayed at the top of the homepage for greater exposure. The Featured section also includes content published by officially certified pet experts, including certified veterinarians and certified pet trainers.

[0134] XI. Data Storage Module

[0135] The data storage module stores various types of data required for system operation. This module adopts a distributed storage architecture and includes three storage units:

[0136] (1) Content data storage unit: stores pet content data, including image files, video files, text content, and metadata information;

[0137] (2) Tag Index Storage Unit: The inverted index structure is used to store the mapping relationship between pet category tags and pet content data, supporting quick retrieval of relevant content based on tags;

[0138] (3) User data storage unit: stores user information, user interest tags, user behavior records, and pet profile information.

[0139] 12. Data Processing Flow

[0140] like Figure 5 As shown, the content data processing flow of this invention is as follows:

[0141] Step S1: Users publish pet content through the client, and the content data collection module receives the pet content data;

[0142] Step S2: The pet classification tag module identifies the breed of pets and generates tags from the pet content data, thus generating pet classification tags;

[0143] Step S3: The content review module performs machine and manual review on the pet content data;

[0144] Step S4: Approved content is stored in the data storage module, while unapproved content is intercepted and the user is notified;

[0145] Step S5: The user profile module updates user interest tags based on user behavior data;

[0146] Step S6: When a user requests to browse content, the content recommendation module generates a personalized recommended content stream based on pet category tags and user interest tags;

[0147] Step S7: The location service module filters local content based on the user's selection;

[0148] Step S8: The interaction statistics module collects the interaction data of the content and updates the overall popularity value;

[0149] Step S9: The Featured Content module filters content that meets the threshold and adds it to the Featured section.

[0150] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention can have various modifications and variations. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A big data management system based on pet classification tags, characterized in that it includes: 1.1 Content data collection module, used to collect pet content data published by users. The pet content data includes text and image data, video data, and nine-grid image data. The pet content data includes pet image information, text description information, publication time information, and publication location information. 1.2 Pet classification tag module, connected to the content data acquisition module, is used to classify and tag the pet content data. It includes a pet breed identification unit and a tag generation unit. The pet breed identification unit identifies pet breed information in the pet content data through image recognition technology. The tag generation unit generates corresponding pet classification tags based on the identification results. The pet classification tags include pet category tags, pet breed tags, and content theme tags. 1.3 The user profile module is used to construct user interest tags based on users' browsing behavior data, interaction behavior data, and attention behavior data. The user interest tags include preferred pet category tags, preferred pet breed tags, and preferred content theme tags. 1.4 Content recommendation module, connected to the pet category tag module and the user profile module, is used to match content based on the pet category tags and the user interest tags to generate a personalized recommendation content stream; 1.5 The geolocation service module is used to obtain the user's geolocation information and filter pet content data within a specified range based on the geolocation information; 1.6 Interactive statistics module, used to collect interactive data of pet content data, including likes, comments, favorites, and feeding data, and calculate the comprehensive popularity value based on the interactive data; 1.7 Data storage module, used to store the pet content data, the pet category tags, the user interest tags, and the interaction data.

2. A big data management system based on pet classification tags according to claim 1, characterized in that, The pet category labels include cats, dogs, and exotic pets, with the exotic pets category including birds and rabbits; 2.1 The content topic tags mentioned include daily record tags, training skill tags, health science tags, and cute pet stories tags.

3. A big data management system based on pet classification tags according to claim 1, characterized in that, The content recommendation module includes a priority recommendation unit, a popularity ranking recommendation unit, and a breed matching recommendation unit. The priority recommendation unit prioritizes displaying pet content data published by authors that the user has followed. The popularity ranking recommendation unit sorts the pet content data according to the overall popularity value. The breed matching recommendation unit matches corresponding pet content data according to the preferred pet breed tags in the user's interest tags.

4. A big data management system based on pet classification tags according to claim 1, characterized in that, The comprehensive popularity value is calculated as follows: the number of likes is multiplied by the first weight coefficient, the number of comments is multiplied by the second weight coefficient, the number of favorites is multiplied by the third weight coefficient, and the number of feeds is multiplied by the fourth weight coefficient. The four products are then added together to obtain the comprehensive popularity value, wherein the first weight coefficient is 1, the second weight coefficient is 2, the third weight coefficient is 3, and the fourth weight coefficient is 5.

5. A big data management system based on pet classification tags according to claim 1, characterized in that, The geolocation service module includes a positioning unit, a distance filtering unit, and a city switching unit. The positioning unit obtains the user's current geographic coordinates. The distance filtering unit filters pet content data within a preset filtering radius, which includes three levels: 5 kilometers, 10 kilometers, and 20 kilometers. The city switching unit allows users to manually switch target cities.

6. A big data management system based on pet classification tags according to claim 1, characterized in that, It also includes a content review module, which comprises a machine review unit and a human review unit. The machine review unit uses artificial intelligence technology to detect illegal content in pet content data, and the human review unit reviews the review results of the machine review unit. Only pet content data that passes the review can enter the content recommendation module for distribution.

7. A big data management system based on pet classification tags according to claim 1, characterized in that, It also includes a pet profile management module, which is used to record the user's pet information, including the first time the pet was owned, the pet breed information, the pet age information, and the vaccination status information. The pet profile management module supports setting the visibility range of the pet information, which includes public visibility, visible to friends, and visible only to the user.

8. A big data management system based on pet classification tags according to claim 1, characterized in that, It also includes a contribution value calculation module, which calculates user contribution value based on users' note-posting, liking, commenting, and sharing behaviors. The user contribution value is used to redeem platform benefits.

9. A big data management system based on pet classification tags according to claim 1, characterized in that, It also includes a featured content module, which is used to filter pet content data whose overall popularity value reaches a preset threshold of 100. Pet content data that reaches the preset threshold will be displayed at the top of the featured section.

10. A big data management system based on pet classification tags according to claim 1, characterized in that, The data storage module adopts a distributed storage architecture, including a content data storage unit, a tag index storage unit, and a user data storage unit. The tag index storage unit uses an inverted index structure to store the mapping relationship between pet category tags and pet content data.

Images