Intelligent work display system based on data labels

Abstract

The invention relates to the field of work display, and discloses an intelligent work display system based on data labels, comprising: a work data acquisition module used for collecting related data of each work, the related data comprising basic information of the work, detailed description of the work, and related pictures or videos of the work; the basic information of the work comprises a work name, an author and creation time; the label generation module is used for preprocessing the data collected by the data acquisition module, inputting the data into a work classification model and outputting a class label of the work; the data storage module is used for storing the work data and the corresponding category labels in a database; and the intelligent display module analyzes the work category preference of the user according to the basic information and the historical operation information of the user, matches the work category preference of the user with the corresponding work category label, and provides personalized work display for the user.

Description

Technical Field

[0001] The present invention relates to the field of work display, and particularly to an intelligent work display system based on data tags. Background Art

[0002] In the current era of digital information explosion, the number of various works has increased exponentially. How to effectively manage and display these works has become an urgent problem to be solved.

[0003] Traditional work display methods mainly rely on manual classification and sorting, which have many drawbacks. On the one hand, the efficiency of manual classification is low, requiring a large amount of time and labor costs. For a vast amount of work data, manually classifying and labeling each one not only involves a huge workload but also is prone to errors and omissions. On the other hand, the subjectivity of manual classification is relatively strong. Different people may have different understandings and classification criteria for works, resulting in inaccurate and inconsistent classification results.

[0004] Although some existing work display systems have introduced certain information management means, most of them simply store and display work information electronically, lacking in-depth analysis and mining of work information; thus, they are unable to make accurate work recommendations and displays according to users' personalized needs, resulting in users often spending a large amount of time and effort screening and searching when looking for works they are interested in, with a poor experience. Summary of the Invention

[0005] The purpose of the present invention is to provide an intelligent work display system based on data tags to solve the technical problem that it is impossible to make accurate work recommendations and displays according to users' personalized needs, resulting in users often spending a large amount of time and effort screening and searching when looking for works they are interested in, with a poor experience.

[0006] The purpose of the present invention can be achieved through the following technical solutions:

[0007] An intelligent work display system based on data tags, comprising:

[0008] A work data collection module for collecting relevant data of each work, where the relevant data includes the basic information of the work, the detailed description of the work, and the relevant pictures of the work; the basic information of the work includes the work name, the author, and the creation time;

[0009] A tag generation module, which preprocesses the data collected by the data collection module and then inputs it into a work classification model to output the category tags of the work;

[0010] A data storage module for storing the work data and the corresponding category tags in a database;

[0011] The intelligent display module analyzes the user's preference for work categories based on the user's basic information and historical operation information, and then matches the user's preference for work categories with the corresponding work category tags to provide personalized work display for the user.

[0012] In a further solution, the work classification model is a composite model of a pre-trained convolutional neural network model and a recurrent neural network model.

[0013] In a further solution, the process of providing personalized work display for the user includes:

[0014] Obtain the basic information of the current user, including the preferred work category and the information of the followed authors;

[0015] Based on the historical operation information of the current user, analyze and obtain the interest degree RK of the current user for each work category; K represents the Kth work category;

[0016] After sorting in descending order of the interest degree RK, select the top N work categories to form a preliminary work recommendation sequence;

[0017] Based on the interest degree RK of the current user for each work category, the popularity coefficient H of each work, and the difference ΔT between the creation time of each work and the current time, comprehensively analyze and obtain the recommendation score PS of each work; S represents the Sth work;

[0018] Re-sort the preliminary work sequence according to the recommendation score PS, and select the top M works to form a work recommendation sequence.

[0019] In a further solution, the historical operation information of the current user includes: the number of views NL, the number of likes Nd, the number of comments Np, and the number of collections NS of the Kth work category; the total number of historical views NL of the current user z , the number of likes Nd z , the number of comments Np z and the number of collections NS z ;

[0020] The expression of the interest degree RK of each work category is:

[0021]

[0022] In the formula, n represents the number of preset time periods, ξ represents the ξth preset time period, θ ξ represents the weight coefficient corresponding to the ξth preset time period, α, β, γ, δ represent weight factors, and α + β + γ + δ = 1, rK ξ represents the interest index of each work category in the ξth preset time period.

[0023] For a further solution, the weight coefficient θ corresponding to the ξ-th preset time period ξ is obtained through the following process:

[0024] Bring the interest index rK of each work category in the ξ-th preset time period and the previous preset time period ξ 、rK ξ-1 into the following formula:

[0025] ΔrK = rK ξ - rK ξ-1 Calculate the difference ΔrK of the interest index of each work category;

[0026] If ΔrK > 0, then when ΔrK ≥ ΔrK o+ at this time, otherwise,

[0027] If ΔrK < 0, then when ΔrK ≤ ΔrK o- at this time, otherwise,

[0028] In the formula, ω is an adjustment coefficient, is the initial weight coefficient.

[0029] For a further solution, the recommended score PS of each work is calculated through the formula: The calculation is obtained; in the formula, ρ1, ρ2, ρ3 represent reference coefficients, and ρ1 + ρ2 + ρ3 = 1.

[0030] For a further solution, the expression of the work heat coefficient H is:

[0031] In the formula, m represents the number of items of work heat-related parameters, ξ j (t) represents the curve of each work heat-related parameter changing with time, t0 and t1 respectively represent the starting time and ending time of the monitoring period, ν j represents the preset proportional coefficient of each work heat-related parameter; j represents the j-th item; the work heat-related parameters include the browsing, liking, commenting, and collecting quantities of the current work.

[0032] For a further solution, the intelligent display module further includes: a user feedback collection unit for collecting user feedback information on the displayed works, and the feedback information includes the recommended satisfaction;

[0033] When the recommended satisfaction is lower than the satisfaction threshold interval, increase the work categories in the preliminary work recommendation sequence; when the recommended satisfaction belongs to the satisfaction threshold interval, keep the work categories in the preliminary work recommendation sequence; when the recommended satisfaction is higher than the satisfaction threshold interval, reduce the work categories in the preliminary work recommendation sequence.

[0034] Advantages of the present invention:

[0035] (1) The present invention collects data through the work data collection module and uses the composite model of the tag generation module for classification, replacing manual classification; compared with the traditional method, it greatly shortens the classification time and reduces the error rate;

[0036] (2) In the present invention, the intelligent display module analyzes based on the user's basic information and historical operation information, combines the work popularity and creation time, provides personalized work display for users, improves the user experience, enhances the user's favor and trust in the platform, and increases the user stickiness and activity. Description of the Drawings

[0037] The present invention will be further described below with reference to the accompanying drawings.

[0038] Figure 1 It is the system logic block diagram of the present invention. Detailed Embodiments

[0039] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art without creative efforts based on the embodiments of the present invention belong to the scope of protection of the present invention.

[0040] Please refer to Figure 1 As shown, the present invention is an intelligent work display system based on data tags, including:

[0041] A work data collection module for collecting relevant data of each work, where the relevant data includes the basic information of the work, the detailed description of the work, and the relevant pictures of the work; the basic information of the work includes the work name, author, and creation time;

[0042] A tag generation module, after preprocessing the data collected by the data collection module, inputs it into the work classification model and outputs the category tags of the work; preprocesses the image and text data to meet the input requirements of the convolutional neural network CNN model and the recurrent neural network RNN model; for example, scales and normalizes the image, and performs word segmentation and word embedding on the text, etc.

[0043] A data storage module for storing the work data and the corresponding category tags in a database;

[0044] The intelligent display module analyzes the user's preference for work categories based on the user's basic information and historical operation information, and then matches the user's preference for work categories with the corresponding work category tags to provide personalized work display for the user.

[0045] In the present invention, in order to solve the problem that before the existing work display, manual classification of works is adopted, the procedure is complex and the classification error rate is high due to subjective factors. Therefore, through the collection and preprocessing of work information, and then the category tags of each work are output through the tag generation module. The category tags are stored in a relational database such as MySQL, Oracle or a non-relational database such as MongoDB, Redis for retrieval and extraction. Obviously, it can make the work classification faster and more accurate, thus creating conditions for providing more preference-compliant work displays for different users in the follow-up, and further achieving the dual effects of improving user satisfaction and work display efficiency.

[0046] The work classification model is a composite model of a pre-trained convolutional neural network model and a recurrent neural network model. In the present invention, the tag generation module uses a composite model of a pre-trained convolutional neural network CNN model and a recurrent neural network RNN model, which fully combines the advantages of CNN in image feature extraction and RNN in text sequence processing, and can more accurately process multi-modal data to generate reasonable category tags. It should be noted that: the pre-trained convolutional neural network model is one of the ResNet, VGG or Inception series models, and the recurrent neural network model is one of the LSTM, GRU models; the composite model uses an early fusion, late fusion or attention mechanism-based fusion method to fuse the output features of the convolutional neural network model and the recurrent neural network model.

[0047] The process of providing personalized work display for the user includes:

[0048] Obtain the basic information of the current user, including the preferred work categories and the information of the authors followed.

[0049] Based on the historical operation information of the current user, analyze and obtain the interest degree RK of the current user for each work category; K represents the Kth work category.

[0050] After sorting in descending order of the interest degree RK, select the top N work categories to form a preliminary work recommendation sequence.

[0051] Based on the interest degree RK of the current user for each work category, the heat coefficient H of each work, and the difference ΔT between the creation time of each work and the current moment, comprehensively analyze and obtain the recommendation score PS of each work; S represents the Sth work.

[0052] Re - sort the preliminary work sequence according to the recommended score PS, and select the top M works to form a work recommendation sequence.

[0053] In the present invention, by obtaining basic user information such as the preferred work categories and the information of followed authors, the preference direction of the user can be initially understood. Then, by combining the analysis of the user's historical operation information to obtain the interest degree RK, the interest degree of the user in different work categories can be more accurately grasped. For example, if a user often browses works of a certain category, the system will assign a higher interest degree to this category and give priority to works of this category in subsequent recommendations. This makes the recommended works more in line with the actual interests of the user, greatly increasing the probability that the user can find interesting works and enhancing the user experience. At the same time, arrange the work categories in descending order of the interest degree RK and select the top N work categories to form a preliminary work recommendation sequence, narrowing the recommendation scope and focusing on the categories that the user is more interested in. After that, comprehensively consider the user interest degree RK, the work popularity coefficient H, and the difference between the creation time of the work and the current time to calculate the recommended score PS, further optimizing the recommendation result. Among them, the popularity coefficient H can reflect the popularity of the work, and newly created works can also obtain appropriate weights in the recommendation through the time difference factor, making the recommended works not only in line with the user's interests but also have a certain degree of popularity and timeliness, improving the quality and efficiency of the recommendation.

[0054] Moreover, when the user can quickly and accurately find works that match their interests, it will increase the user's favor and trust in the platform, and thus the user will be more willing to use the platform frequently, improving the user stickiness and activity. For example, if a user can always find new works they are interested in on the platform, they will be more inclined to use the platform in the long - term to obtain work information and perform related operations, which is of positive significance for the long - term stable development of the platform. Similarly, this personalized recommendation method avoids blind recommendations and reduces the situation where users browse a large number of uninteresting works, enabling the platform's resources such as server resources and storage resources to be more effectively utilized. The platform can focus on recommending works that the user may be interested in, improving the utilization efficiency of resources and reducing operating costs.

[0055] The historical operation information of the current user includes: the browsing times NL, the number of likes Nd, the number of comments Np, and the number of collections NS of the K - th type of works; the total historical browsing times NL of the current user z 、the number of likes Nd z 、the number of comments Np z and the number of collections NS z ;

[0056] The expression of the interest degree RK for each type of work is:

[0057]

[0058] In the formula, n represents the number of preset time periods, ξ represents the ξ-th preset time period, and θ ξ represents the weight coefficient corresponding to the ξ-th preset time period, and α, β, γ, and δ represent weight factors, which are determined comprehensively based on historical data and empirical data, and α + β + γ + δ = 1, rK ξ represents the interest index of each work category in the ξ-th preset time period.

[0059] The present invention provides a method for obtaining the interest degree RK of each work category, and calculates the interest degree RK of each work category through calculation; the above method comprehensively considers the browsing, liking, commenting, and collecting behaviors of users on different work categories within different preset time periods, and dynamically adjusts the calculation of the interest degree of users on various works through weight factors and preset time period weight coefficients; it more precisely reflects the change of user interests over time, makes the calculation result of the interest degree more accurate, and thus provides more accurate data support for personalized recommendation.

[0060] The weight coefficient θ corresponding to the ξ-th preset time period ξ is obtained through the following process:

[0061] Substitute the interest index rK of each work category in the ξ-th preset time period and the previous preset time period ξ 、rK ξ-1 into the following formula:

[0062] ΔrK = rK ξ -rK ξ-1 to calculate the difference ΔrK of the interest index of each work category;

[0063] If ΔrK > 0, then when ΔrK ≥ ΔrK o+ at this time, otherwise,

[0064] If ΔrK < 0, then when ΔrK ≤ ΔrK o- at this time, otherwise,

[0065] In the formula, ω is an adjustment coefficient, which is determined through analysis of historical data and empirical data, is the initial weight coefficient, ΔrK o+ 、ΔrK o- are preset warning values.

[0066] The present invention provides a method for obtaining the weight coefficient θ corresponding to each preset time period ξ , specifically, according to ΔrK = rK ξ -rK ξ-1Calculate the difference ΔrK of the interest indicators for each work category, which represents the dynamic adjustment of the weight coefficient based on the change of the interest indicators of the work category in adjacent preset time periods, and can capture the fluctuations of user interests in a timely manner; when the user's interest rises or falls significantly, the weight is adjusted accordingly, so that the calculation of the user's interest degree can better reflect their latest interest tendency, and further optimize the personalized recommendation effect.

[0067] The recommended score PS of each work is calculated by the formula: It is calculated; in the formula, ρ1, ρ2, and ρ3 represent reference coefficients, which are determined based on historical data analysis, and ρ1 + ρ2 + ρ3 = 1.

[0068] In the present invention, RK represents the user's interest degree in the Kth work category, which reflects the user's personalized preference. The higher the RK value, the stronger the user's interest in the works of this category; H is the work heat coefficient, which is used to measure the popularity of the work among the public. The larger the H value, the higher the heat of the work; ΔT represents the difference between the creation time of the work and the current moment. The larger ΔT is, the longer the work is from the current time; this formula takes these three factors into consideration and comprehensively weighs the user's personalized needs, the popularity of the work, and timeliness; multiplying the work heat coefficient H by the reciprocal of the creation time difference ΔT means that the higher the work heat and the newer the creation time, the greater the score increase in the recommended score; plus the user's interest degree RK in the work category, the finally obtained recommended score PS can more comprehensively reflect the recommended value of the work for the user; thus, the works can be sorted according to the PS value, and the works with high PS values are recommended first, ensuring that the recommended works not only match the user's interests, but also have a certain degree of heat and freshness, optimizing the recommendation results and enhancing the attractiveness and value of the recommended works to the user.

[0069] The expression of the work heat coefficient H is:

[0070] In the formula, m represents the number of terms of the work heat-related parameters, ξ j (t) represents the curve of each work heat-related parameter changing with time, t0 and t1 respectively represent the starting moment and the ending moment of the monitoring period, ν j represents the preset proportional coefficient of each work heat-related parameter, which is comprehensively determined based on historical data; j represents the jth item; the work heat-related parameters include but are not limited to the browsing, liking, commenting, and collecting quantities of the current work.

[0071] In the present invention, by comprehensively considering the parameters closely related to the work heat, such as the browsing, liking, commenting, and collecting quantities of the work, it is possible to measure the degree of attention and popularity of the work among users on the platform more objectively and comprehensively; compared with simply judging the work heat based on a single indicator, the comprehensive calculation method is more accurate and avoids the deviation of the work heat evaluation caused by the limitations of a single indicator; by The cumulative change amount of each parameter within the monitoring period is summed and averaged, so that the work heat coefficient H can reflect the dynamic change of the work heat over time, thereby capturing the fluctuation of the work heat, making timely evaluation and adjustment for the works with rising or falling heat, and providing more timely data support for work recommendation.

[0072] The intelligent display module further includes: a user feedback collection unit for collecting user feedback information on the displayed works. The feedback information includes recommendation satisfaction, which is reflected by user scoring. The score is set from 1 to 10, and the higher the score, the higher the recommendation satisfaction.

[0073] When the recommendation satisfaction is lower than the satisfaction threshold range, increase the work categories in the preliminary work recommendation sequence; when the recommendation satisfaction belongs to the satisfaction threshold range, keep the work categories in the preliminary work recommendation sequence; when the recommendation satisfaction is higher than the satisfaction threshold range, reduce the work categories in the preliminary work recommendation sequence.

[0074] It should be noted that: the calculation formulas and each parameter participating in the operation in the present invention are all pre-dimensionless processed, and the process of dimensionless processing is well-known in the industry and will not be described here.

[0075] The above has described a detailed embodiment of the present invention, but the content described is only the preferred embodiment of the present invention and cannot be considered as used to limit the implementation scope of the present invention. All equivalent changes and improvements made according to the application scope of the present invention should still fall within the patent coverage scope of the present invention.

Claims

1. An intelligent work display system based on data tags, characterized in that: include: The work data collection module is used to collect relevant data of each work, wherein the relevant data includes basic information of the work, detailed description of the work, and relevant pictures of the work; the basic information of the work includes the work name, author, and creation time; A label generation module pre-processes the data collected by the data acquisition module, inputs the data into the work classification model, and outputs the category label of the work; A data storage module, used to store work data and corresponding category labels in a database; The intelligent display module analyzes the user's work category preferences based on the user's basic information and historical operation information, and then matches the user's work category preferences with the corresponding work category labels to provide users with personalized work display.

2. The intelligent work display system based on data tags according to claim 1 is characterized in that: The work classification model is a composite model of a pre-trained convolutional neural network model and a recurrent neural network model.

3. The intelligent work display system based on data tags according to claim 2 is characterized in that: The process of providing users with personalized work display includes: Get the basic information of the current user, including favorite work categories and authors they follow; Based on the historical operation information of the current user, the current user's interest in each work category RK is analyzed; K represents the Kth work category; After arranging in descending order of interest RK, select the top N categories of works to form a preliminary work recommendation sequence; Based on the current user's interest in each work category RK, the popularity coefficient H of each work, and the difference ΔT between the creation time of each work and the current time, a comprehensive analysis is performed to obtain the recommendation score PS of each work; S represents the Sth work; The preliminary works sequence is reordered according to the recommendation score PS, and the works in the top M positions are selected to form a work recommendation sequence.

4. The intelligent work display system based on data tags according to claim 3 is characterized in that: The historical operation information of the current user includes: the number of views NL, the number of likes Nd, the number of comments Np and the number of favorites NS of the Kth work category; the total number of historical views NL of the current user z 、Nd number of likes z 、Np number of comments z and number of collections NS z ; The expression of interest RK of each work category is: Where n represents the number of preset time periods, ξ represents the ξth preset time period, θ ξ represents the weight coefficient corresponding to the ξth preset time period, α, β, γ, δ represent weight factors, and α+β+γ+δ=1, rK ξ Represents the interest index of each work category in the ξ-th preset period.

5. The intelligent work display system based on data tags according to claim 4 is characterized in that: The weight coefficient θ corresponding to the ξ-th preset time period ξ Obtained through the following process: Compare the interest index rK of each work category in the ξth preset period with that in the previous preset period ξ , rK ξ-1 Substitute the following formula: ΔrK=rK ξ -rK ξ-1 Calculate the difference ΔrK of the interest index of each work category; If ΔrK>0, then when ΔrK≥ΔrK o+ hour, otherwise, If ΔrK<0, then when ΔrK≤ΔrK o- hour, otherwise, In the formula, ω is the adjustment coefficient, is the initial weight coefficient, ΔrK o+ , ΔrK o- It is the preset warning value.

6. The intelligent work display system based on data tags according to claim 3 is characterized in that: The recommended score PS of each work is calculated by the formula: Calculated; where ρ1, ρ2, ρ3 represent reference coefficients, and ρ1+ρ2+ρ3=1.

7. The intelligent work display system based on data tags according to claim 6 is characterized in that: The expression of the work heat coefficient H is: In the formula, m represents the number of parameters related to the popularity of the work, ξ j (t) represents the time-varying curve of the popularity-related parameters of each work, t0 and t1 represent the start and end time of the monitoring period, respectively, and ν j Represents the preset proportional coefficient of the popularity-related parameters of each work; j represents the jth item.

8. The intelligent work display system based on data tags according to claim 3 is characterized in that: The intelligent display module also includes: a user feedback collection unit, which is used to collect user feedback information on the displayed works, and the feedback information includes recommendation satisfaction; When the recommendation satisfaction is lower than the satisfaction threshold interval, the categories of works with heavy weight in the preliminary work recommendation sequence are increased; when the recommendation satisfaction is within the satisfaction threshold interval, the categories of works with heavy weight in the preliminary work recommendation sequence are maintained; when the recommendation satisfaction is higher than the satisfaction threshold interval, the categories of works with heavy weight in the preliminary work recommendation sequence are reduced.

Images