Digital new media youth innovation information management method based on big data

Abstract

The invention discloses a digital new media youth innovation information management method based on big data, which relates to the technical field of innovation information management. The invention performs preference analysis and prediction on innovation content, performs similarity analysis on two adjacent interest preferences, and obtains corresponding similarity data; generates preference variation from similarity data, extracts innovation topics when the preference variation is lower than expected, analyzes and obtains central values ​​of several innovation topics, and screens out key topics from several innovation topics by the central values; collects innovation content dissemination feedback data containing key topics, generates dissemination degree from the dissemination feedback data, screens out target projects in the innovation content by the dissemination degree, and obtains several path classes after classifying the innovation paths of the generated target projects; screens several innovation contents according to key topics, and processes key topics in a targeted manner, thereby improving management efficiency.

Description

Technical Field

[0001] The present invention relates to the technical field of innovative information management, and in particular to a method for managing innovative information of young people using digital new media based on big data. Background Art

[0002] Youth innovation information management is a comprehensive process, which involves systematically collecting, organizing, storing, analyzing and efficiently disseminating youth’s innovative ideas, project results and activity information in multiple fields such as scientific exploration, artistic creation, and social practice. It aims to stimulate and cultivate young people’s innovative thinking and practical operation ability, promote the sharing and in-depth communication of creative resources, and build a solid information support platform for the effective discovery and comprehensive cultivation of young innovative talents, thereby promoting the overall improvement of young people’s innovation ability.

[0003] The application publication number CN118568637A discloses an intelligent management system for scientific and technological innovation experimental data, which relates to the field of experimental data technology, including an experimental data acquisition module, a related type table generation module, a related display module and a data anomaly type output module; through the coordinated use of the experimental data acquisition module, the related type table generation module, the related display module and the data anomaly identification module, the newly entered or modified experimental data is detected in real time, and compared with the theoretical value obtained by the calculation function, the system can quickly identify potential data anomalies, avoid data errors or omissions when manually entering or modifying data, and reduce the impact of human factors on data quality. Through automated data anomaly detection, it can quickly identify abnormal data types and mark and output them, reducing the time spent by experimenters on data verification and error troubleshooting.

[0004] Combined with the above application and the contents of the prior art:

[0005] Teenagers have quick thinking, rich creative thinking, and usually have many innovative ideas. Today's new media is very developed, and the instant sparks of innovation are usually recorded by teenagers on new media. When managing teenagers' innovation information and data, these data are usually collected and recorded for subsequent reference.

[0006] In the existing management methods of innovation information and data, after obtaining innovative content, the obtained content is usually pre-processed and classified, and then corresponding innovation resource support is provided for each type of innovative content, such as professional tutoring and technical support. However, considering that new media content is rich and the speed of dissemination is relatively fast, many young people's innovative content also has great similarities. If support is given in the management of innovative content, it will be difficult to achieve when the innovation support resources are limited. Moreover, the innovation support resources provided may be difficult to match with youth innovation projects.

[0007] To this end, the present invention provides a digital new media youth innovation information management method based on big data. Summary of the invention

[0008] 1. Technical issues to be resolved

[0009] In view of the deficiencies in the prior art, the present invention provides a digital new media youth innovation information management method based on big data, which performs preference analysis and prediction on innovative content, performs similarity analysis on two adjacent interest preferences, and obtains corresponding similarity data; generates preference change from the similarity data, extracts innovative topics when the preference change is lower than expected, analyzes and obtains the central values ​​of several innovative topics, and screens out key topics from several innovative topics by the central values; collects innovative content dissemination feedback data containing key topics, generates dissemination degrees from the dissemination feedback data, screens out target projects in the innovative content by the dissemination degrees, and obtains several path classes after classifying the innovative paths of the generated target projects; screens several innovative contents according to the key topics, performs targeted processing on the key topics, and improves management efficiency; thereby solving the technical problems recorded in the background technology.

[0010] (II) Technical solution

[0011] To achieve the above objectives, the present invention is implemented through the following technical solutions: a digital new media youth innovation information management method based on big data, including collecting the innovation content of target teenagers on the new media platform, analyzing and predicting the preference of the innovation content, analyzing the similarity between two adjacent interest preferences, and obtaining corresponding similarity data;

[0012] Calculate the preference change of interest preference from similarity data , in the preference change When the performance is lower than expected, we extract innovation themes and analyze and obtain the central values ​​of several innovation themes. , by the center value Screen out key themes from several innovation themes;

[0013] Collect feedback data on the dissemination of innovative content containing key topics, and generate dissemination scores based on the dissemination feedback data , by the propagation Filter out target projects from the innovation content, classify the innovation paths of the generated target projects, obtain several path classes, and send push instructions to the outside;

[0014] Push innovation support resources for each path type, record and obtain corresponding execution status data, and generate project completion from execution feedback data. If the project completion exceeds the completion threshold, the corresponding innovation path will be used as a path to be adjusted;

[0015] After collecting the path adjustment feedback data of the target adolescents, the optimization features are extracted. The innovation path optimization knowledge graph outputs an optimization plan for optimizing the innovation path based on the optimization features. When the feasibility of the optimization plan is not sufficient, innovation support resources are pushed to the optimized innovation path.

[0016] Furthermore, we collect various innovative contents uploaded by the target teenagers, including video images and text content, use NLP algorithms to perform content analysis on the text content, and extract the interests, hobbies and emotional tendencies of the target teenagers; analyze the content themes, colors and character activities in the video images, and summarize them to generate an innovative activity data set.

[0017] Furthermore, the pre-trained K-Means algorithm extracts the interest preference data of the target teenagers from the innovation activity data and marks them with timestamps; several consecutive interest preferences are sorted along the timeline to obtain the corresponding preference columns, and after predicting the interest preferences using time series analysis, the obtained prediction results are added to the preference columns.

[0018] Furthermore, the preference change of interest preference is calculated based on the continuous similarity data , if the preference change If the change degree is lower than the threshold, a topic extraction instruction is issued to the outside; the preference change degree The generation method is as follows:

[0019] ;

[0020] in, is the i-th similarity, is the mean of similarity, is the standard deviation of similarity; weight coefficient: , ,and .

[0021] Furthermore, the innovative content output by the target youth in the current stage is summarized, including text data and multimedia content data; after receiving the topic extraction instruction, the pre-trained LDA algorithm is used to identify the innovative topics in the innovative content data, and the trained deep learning model adds corresponding labels to the innovative content according to the extracted innovative topics;

[0022] After arranging several innovation themes, a theme network diagram is constructed, and the corresponding similarity data is obtained after similarity analysis is performed between each innovation theme.

[0023] Furthermore, the central values ​​of several innovation topics are calculated by obtaining similarity data. , if the center value Exceeding the central threshold, the innovation theme is considered as a key theme; the central value The generation method is as follows:

[0024] ;

[0025] Where: is the number of innovation themes, is the variance of similarity, and is the weight coefficient, weight coefficient: , ,and ; For the theme and The similarity between is the corresponding time interval, is the time decay function, is the covariance of similarity and time interval.

[0026] Furthermore, when the innovative content containing key themes is in a public dissemination state, the corresponding dissemination feedback data, including the number of forwarding, the number of views, and the distribution range of viewers, are collected and summarized as a dissemination feedback data set, and the dissemination degree is generated from the dissemination feedback data in the dissemination feedback data set. , if the propagation If the number is higher than the preset propagation threshold, it will be targeted as a project;

[0027] After the target project determines the identity information of the corresponding target youth, the learning and innovation data of the target youth are collected, including innovation content, learning content and knowledge and technology reserves, and the current innovation resources and interest preference data are summarized to generate an innovative learning data set.

[0028] Furthermore, under dimensionless conditions, the propagation degree is generated from the propagation feedback data The way is as follows:

[0029] ;

[0030] Where: , is the weight coefficient of time t and spatial position x, is the time-dependent weight coefficient, is the number of likes at time t; is the time attenuation coefficient, is the second-order rate of change of the forwarding amount in time and space, is the second-order rate of change of page views in time and space, For the propagation cycle.

[0031] Furthermore, using the innovation learning data as input, a pre-trained path planning algorithm is used to plan the innovation path for the target youth, generating a personalized innovation path that is suitable for the target youth;

[0032] After aggregating several personalized innovation paths, use a pre-trained classification algorithm to classify the personalized innovation paths, obtain several path classes, and send push instructions to the outside.

[0033] Furthermore, several innovation support resources are summarized in advance to generate an innovation resource library; after receiving the push instruction, the innovation support resources are pushed to each path type based on the collaborative filtering algorithm to complete the push, and the push feedback data is collected, and the current resource push strategy is iteratively optimized based on the collected feedback data.

[0034] Furthermore, the execution status data of the innovation projects is recorded, including the execution progress of the target projects and the target achievement rate;

[0035] Taking the execution status data as input, the trained completion evaluation model is used to output the project completion. If the obtained completion exceeds the completion threshold, the corresponding innovation path is taken as the path to be adjusted, and adjustment instructions are issued to the outside.

[0036] Further, after receiving the adjustment instruction, after the path to be optimized is matched with the corresponding target youth, the target youth's innovation content dissemination data, resource support data, target project execution data and project feedback data are collected and summarized as path adjustment feedback data;

[0037] Conduct principal component analysis on path adjustment feedback data and completion, extract key factors that affect the completion of target project execution, use youth innovation content optimization as the target word, and pre-construct an innovation path optimization knowledge graph.

[0038] Furthermore, the trained solution prediction model is used to predict the optimization solution, obtain corresponding prediction data, and determine the expected completion degree of the optimized innovation path based on the prediction data;

[0039] If the expected completion rate still does not exceed expectations, after executing the optimization plan to optimize the path to be adjusted, innovation support resources are pushed to the optimized innovation path based on the collaborative filtering algorithm.

[0040] (III) Beneficial effects

[0041] The present invention provides a digital new media youth innovation information management method based on big data, which has the following beneficial effects:

[0042] 1. By extracting and classifying features from the data, determining young people's interest preferences in innovative content from the innovative content data, and making targeted recommendations and matches in the management of post-innovative content, resource support can be achieved.

[0043] 2. Constructing the degree of preference change from preference change data , by evaluating and judging the stability of teenagers' interest preferences, several teenagers can be screened to determine whether corresponding innovative content needs to be selected; through content recognition, theme extraction and summary generation, the classification of various innovative contents can also facilitate quick acquisition of the general meaning of the innovative content.

[0044] 3. Based on the center value Filter out key topics from several innovation topics, and filter several innovation contents based on the key topics. When innovation support and resources are limited, target some key topics, which can improve management efficiency when managing innovation contents.

[0045] 4. Based on the generated spread Evaluate several innovations that cover key topics and obtain the most popular ones, which can be chosen to be prioritized when providing resource support; generate corresponding personalized innovation paths for target youth through the collected relevant data, classify the innovation content according to the innovation path, and choose different innovation support for different innovation paths.

[0046] 5. Provide targeted resource support when the current innovation content is very valuable so that innovation projects can make more progress; track the execution progress of innovation projects, and generate corresponding project completion from analytical data; evaluate the completion of innovation projects based on project completion, and make timely adjustments when the completion is lower than expected to ensure that innovation projects can be completed in a timely manner.

[0047] 6. Match the corresponding optimization plan for the current personalized innovation path based on key factors. Before or after optimizing the innovation path, continue to provide resource support for the corresponding innovation content to accelerate or ensure the completion of the innovation content. In this way, when managing the innovation content, in addition to improving management efficiency, it can also improve the targeted management and reduce the waste of innovation projects and innovation resources. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 It is a schematic diagram of the flow chart of the innovative information management method of the present invention. DETAILED DESCRIPTION

[0049] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0050] See also Figure 1 ,The present invention provides a digital new media youth innovation information management method based on big data, including, step 1, collecting the innovation content of target youth on the new media platform, performing preference analysis and prediction on the innovation content, analyzing the similarity between two adjacent interest preferences, and obtaining corresponding similarity data;

[0051] The step 1 includes the following contents:

[0052] Step 101: collect various innovative contents uploaded by target teenagers on the new media platform, including video images and text contents, use NLP algorithms to perform content analysis on the text contents, and extract the interests, hobbies and emotional tendencies of the target teenagers;

[0053] Use pre-trained image analysis and recognition algorithms and theme extraction algorithms to analyze the content, theme, color, and character activities in video images, and then aggregate them to generate an innovative activity data set;

[0054] Step 102: extracting the interest preference data of target teenagers from the innovation activity data using a pre-trained K-Means algorithm, and marking the interest preference with a timestamp;

[0055] When in use, based on continuous data collection, by extracting and classifying data features, the interest preferences of young people in innovative content can be determined from the innovative content data, and targeted recommendations and matching can be made in the management of post-innovative content to achieve resource support.

[0056] Sort several consecutive interest preferences along the time axis to obtain the corresponding preference column, use time series analysis to predict the interest preferences, add the obtained prediction results to the preference column, perform similarity analysis on two adjacent interest preferences, and obtain the corresponding similarity data;

[0057] When using, combine the contents in steps 101 to 102:

[0058] As a further content, the interests and hobbies of teenagers are predicted based on the existing data, and the prediction and analysis of their interest preferences when they are innovating are completed, so that changes in interest preferences can be dealt with in advance.

[0059] In the existing management methods of innovation information and data, after obtaining innovative content, the obtained content is usually pre-processed and classified, and then corresponding innovation resource support is provided for each type of innovative content, such as professional tutoring and technical support. However, considering that new media content is rich and the speed of dissemination is relatively fast, many young people's innovative content also has great similarities. If support is given in the management of innovative content, it will be difficult to achieve when the innovation support resources are limited. Moreover, the innovation support resources provided may be difficult to match with youth innovation projects.

[0060] Step 2: Calculate the preference change of interest preference based on similarity data , in the preference change When the performance is lower than expected, we extract innovation themes and analyze and obtain the central values ​​of several innovation themes. , by the center value Screen out key themes from several innovation themes;

[0061] The step 2 includes the following contents:

[0062] Step 201: Calculate the preference change of interest preference based on continuous similarity data , with the generated preference variation Evaluate and describe the extent of changes in interest preferences. The generation method is as follows:

[0063] ;

[0064] in, is the i-th similarity, is the mean of similarity, is the standard deviation of similarity; weight coefficient: , ,and ;The weight coefficient can be obtained by referring to the hierarchical analysis method;

[0065] According to historical data and management expectations of interest preferences, a threshold of actual change degree is set in advance; if the obtained preference change degree If the change rate is lower than the preset threshold, it means that the innovation content of the target youth has not changed much in the recent period, and further topic extraction can be performed. At this time, a topic extraction instruction is issued to the outside.

[0066] When used, based on the prediction of interest preferences, similarity analysis is performed on interest preferences, and the degree of preference change is constructed from the preference change data. , according to the preference change Evaluate and judge the stability of young people's interest preferences, so as to screen several young people and determine whether corresponding innovative content needs to be selected;

[0067] Step 202: Summarize the innovative content output by the target youth in the current stage, including text data and multimedia content data; after receiving the topic extraction instruction, use the pre-trained LDA algorithm to identify the innovative topics and associations in the innovative content data, and use the trained deep learning model to add corresponding labels to the innovative content according to the extracted innovative topics;

[0068] As a further content, on the basis of collecting various innovative contents, through content identification and theme extraction, a summary can be further generated on this basis, which can realize the classification of various innovative contents, and at the same time, it is also convenient to quickly obtain the general meaning of the innovative contents;

[0069] Step 203: Arrange several innovation themes and construct a theme network diagram. Perform similarity analysis on the innovation themes and obtain corresponding similarity data. Under dimensionless conditions, calculate the central values ​​of several innovation themes through the obtained similarity data. , as follows:

[0070] ;

[0071] Where: is the number of innovation themes, is the variance of similarity, and is the weight coefficient, weight coefficient: , ,and ;The weight coefficient can be obtained by referring to the hierarchical analysis method; For the theme and The similarity between is the corresponding time interval, is the time decay function, is the covariance between similarity and time interval;

[0072] Pre-set center thresholds based on historical data and management expectations for innovative content;

[0073] If the center value is obtained If the centrality threshold is exceeded, it means that the centrality of the current innovation theme is high, and it can be used as a key theme, and the innovation content containing the corresponding key theme can be supported;

[0074] When using, combine the contents in steps 201 to 203:

[0075] As a further content, based on the extraction of topics for each innovative content, a central value can be generated based on the similarity and time interval between each topic. By screening out key themes from several innovation themes, we can screen several innovation contents based on key themes when there are many innovation contents, and deal with some key themes in a targeted manner when innovation support and resources are limited, which can improve management efficiency when managing innovation contents.

[0076] Step 3: Collect feedback data on the dissemination of innovative content containing key themes, and generate dissemination degrees from the dissemination feedback data , by the propagation Filter out target projects from the innovation content, classify the innovation paths of the generated target projects, obtain several path classes, and send push instructions to the outside;

[0077] The step three includes the following contents:

[0078] Step 301: When the innovative content containing the key theme is in a public dissemination state, corresponding dissemination feedback data is collected, including the forwarding volume, the number of views, and the distribution range of the viewers, and the data is summarized as a dissemination feedback data set;

[0079] Generate propagation degree from propagation feedback data , based on the degree of spread Evaluate the current dissemination status of innovative content and the degree of dissemination The generation method is as follows:

[0080] ;

[0081] Where: , is the weight coefficient of time t and spatial position x, is the time-dependent weight coefficient, and the weight value falls between 0 and 1. is the number of likes at time t; is the time attenuation coefficient, ranging from 0 to 1. is the second-order rate of change of the forwarding amount in time and space, is the second-order rate of change of page views in time and space, e can be 2.713, for the propagation cycle;

[0082] If the spread If the number is higher than the preset dissemination threshold, it means that the innovation project has received a lot of attention and may need more innovation resources. It should be taken as a target project. At this time, a classification instruction should be issued to the outside world.

[0083] When used, as a further content, after screening the innovative content by key themes, the dissemination data of the innovative content is collected, and the dissemination feedback data is used to generate the dissemination degree. , based on the generated propagation Evaluate several innovations covering key themes and obtain the most popular ones. When it is necessary to manage the innovation content, you can choose to focus on them when providing resource support.

[0084] Step 302: After the target project determines the identity information of the corresponding target youth, collect the learning and innovation data of the target youth, including innovation content, learning content, knowledge and technology reserves, etc., and generate an innovation learning data set after summarizing the current innovation resources and interest preference data;

[0085] Step 303: using the innovation learning data as input, using the pre-trained path planning algorithm to plan the innovation path for the target youth, and generating a personalized innovation path suitable for the target youth;

[0086] After aggregating several personalized innovation paths, use the pre-trained classification algorithm to classify the personalized innovation paths, obtain several path classes, and send push instructions to the outside;

[0087] When using, combine the contents in steps 301 to 303:

[0088] As a further content, after completing the selection among several innovative contents, the corresponding personalized innovation paths can be generated for the target teenagers through the collected relevant data. The innovative content and its changing process can be described and described. Combined with the theme and summary of the innovative content, the innovative content can be classified according to the innovative path, and different innovation supports can be selected for different innovative paths.

[0089] Step 4: Push innovation support resources for each path type, record the corresponding execution status data, and generate project completion from the execution feedback data. If the project completion exceeds the completion threshold, the corresponding innovation path will be used as a path to be adjusted;

[0090] The step 4 includes the following contents:

[0091] Step 401: Aggregate several innovation support resources, such as professional knowledge guidance and innovation resource support, to generate an innovation resource library in advance; after receiving the push instruction, push the innovation support resources for each path type based on the collaborative filtering algorithm to complete the push, collect push feedback data, and iteratively optimize the current resource push strategy based on the collected feedback data until the push process is completed;

[0092] When in use, the recommendation of innovative resources is completed on the basis of further classification of various innovative contents. When the current innovative content is very valuable, targeted resource support is provided to enable the innovative project to make more progress.

[0093] Step 402: When the innovation projects of each path type are in a continuous development state, the execution status of the innovation projects is recorded and updated in real time, and corresponding execution status data is obtained, including the execution progress and target achievement rate of the target projects;

[0094] The neural network is trained with the labeled sample data to obtain the trained completion evaluation model; the execution status data is used as input, and the trained completion evaluation model is used to output the project completion, and the execution status of the target project is evaluated based on the project completion, and the innovation management expectations of the target youth are set in advance with the completion threshold;

[0095] If the degree of completion obtained exceeds the completion threshold, the corresponding innovation path will be taken as the path to be adjusted; this indicates that the effectiveness of the innovation path of the target youth may be insufficient, which may be due to the lack of resource support and guidance, and needs to be dealt with in a targeted manner. At this time, an adjustment instruction is issued to the outside;

[0096] When using, combine the contents in steps 401 to 402:

[0097] As a further content, on the basis of selectively completing resource support, guide the target young people to implement innovative content and corresponding innovative projects, track the implementation progress of the innovative projects, and generate the corresponding project completion degree by analyzing the data. Evaluate the completion of the innovative projects based on the project completion degree, and make timely adjustments when the completion degree is lower than expected to ensure that the innovative projects can be completed in time.

[0098] Step 5: After collecting the path adjustment feedback data of the target youth, the optimization features are extracted, and the innovation path optimization knowledge graph outputs the optimization plan for optimizing the innovation path based on the optimization features. When the feasibility of the optimization plan is insufficient, innovation support resources are pushed for the optimized innovation path;

[0099] The five include the following:

[0100] Step 501: after receiving the adjustment instruction, after the path to be optimized matches the corresponding target youth, the target youth's innovative content dissemination data, resource support data, target project execution data and project feedback data are collected and summarized as path adjustment feedback data;

[0101] Conduct principal component analysis on the path adjustment feedback data and completion degree to extract the key factors that affect the completion degree of the target project execution, and obtain the corresponding optimization features after feature extraction;

[0102] Step 502: Taking youth innovation content optimization as the target word, after deep retrieval and entity relationship building, pre-construct an innovation path optimization knowledge graph; based on the correspondence between the optimization features and the path optimization scheme, output an optimization scheme for optimizing the innovation path from the innovation path optimization knowledge graph;

[0103] When in use, as further feedback, as feedback on the personalized innovation path, when the completion status of the current innovation project is inconsistent with expectations, the key factors affecting the completion of the current project can be identified through principal component analysis, and the corresponding optimization plan can be matched for the current personalized innovation path based on the key factors, so as to optimize the current innovation path in a targeted manner.

[0104] Step 503: train a convolutional neural network with the labeled sample data to obtain a trained solution prediction model; use the trained solution prediction model to predict the optimization solution, obtain corresponding prediction data, and determine the expected degree of completion of the optimized innovation path based on the prediction data; if the expected degree of completion still does not exceed expectations, execute the optimization solution to optimize the path to be adjusted, and push innovation support resources for the optimized innovation path based on the collaborative filtering algorithm.

[0105] When using, combine the contents in steps 501 to 503:

[0106] As feedback on the innovation path, before or after the optimization of the innovation path, continue to provide resource support for the corresponding innovation content to accelerate or ensure the completion of the innovation content. Therefore, when managing the innovation content, in addition to improving management efficiency, it can also improve the targeted management and reduce the waste of innovation projects and innovation resources.

[0107] With the goal of optimizing youth innovation content, the specific steps for pre-constructing an innovation path optimization knowledge graph can be referred to as follows:

[0108] Clarify goals and needs - Goal definition: First, clarify the goal of building a knowledge graph, that is, to optimize youth innovation content and enhance youth's scientific and technological innovation practice capabilities. Demand analysis: Analyze the difficulties, challenges and bottlenecks that young people may encounter in the process of scientific and technological innovation, and how to provide effective solutions and resources through knowledge graphs.

[0109] Deep retrieval and entity recognition - data collection: Use search engines, academic databases, patent databases and other resources to collect multimodal data such as text, images, videos and other data related to youth science and technology innovation. Entity recognition: Identify entities related to youth innovation content from the collected data, such as innovation methods, innovation tools, innovation cases, innovation challenges, etc. Relationship extraction: Use natural language processing and machine learning technology to extract the relationship between entities from text data, such as which innovation challenge a certain innovation method is suitable for solving.

[0110] Entity relationship building and knowledge modeling - Relationship definition: clarify the relationship types between entities, such as "innovation method - solution - innovation challenge", "innovation tool - support - innovation process", etc. Knowledge modeling: build a model of the knowledge graph based on the identified entities and extracted relationships. This includes defining the attributes of entities, the attributes of relationships, and the way they are related. Model layer design: design the model layer of the knowledge graph, that is, define the types and attributes of entities and relationships, as well as the constraints and rules between them.

[0111] Knowledge graph construction and optimization—data integration: Integrate the identified entities and extracted relationships into the knowledge graph to form a preliminary knowledge graph. Knowledge fusion: Fuse the integrated knowledge to eliminate contradictions and ambiguities and ensure the accuracy and consistency of knowledge. Knowledge reasoning: Use logical reasoning, graph-based reasoning, deep learning and other technologies to complete and reason about the relationships in the knowledge graph and discover potential knowledge associations and innovation paths. Quality assessment: Conduct quality assessment on the constructed knowledge graph, including aspects such as accuracy, completeness, consistency and availability of knowledge. Optimization and iteration: Based on the results of the quality assessment, optimize and iterate the knowledge graph to continuously improve its quality and effectiveness.

[0112] Innovation path optimization knowledge graph application - Personalized recommendation: Based on the innovation needs and interests of young people, use the knowledge graph to provide personalized recommendations for innovation methods and tools. Innovation case sharing: Use the knowledge graph to display successful innovation cases and solutions to provide reference and inspiration for young people. Innovation challenge solving: Use the knowledge and relationships in the knowledge graph to help young people solve the challenges and problems encountered in the innovation process. Innovation community construction: Build an innovation community based on the knowledge graph to promote communication and cooperation among young people and jointly promote the optimization and development of innovation content.

[0113] The analytic hierarchy process (AHP) is a structured and systematic decision support technology. It decomposes complex multi-criteria decision problems into a series of interrelated and hierarchical sub-problems or factors, uses expert judgment to compare and quantitatively evaluate the importance of factors at each level, and finally obtains the relative advantages and disadvantages of each decision plan through synthetic calculation, providing decision makers with a clear and quantitative basis for decision making. This method combines qualitative and quantitative analysis, effectively improving the scientificity and accuracy of decision-making.

[0114] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0115] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0116] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only some logical function divisions. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0117] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0118] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A digital new media youth innovation information management method based on big data, characterized by: include, Collect innovative content of target teenagers on new media platforms, analyze and predict the preferences of innovative content, analyze the similarity between two adjacent interest preferences, and obtain corresponding similarity data; Calculate the preference change of interest preference from similarity data , in the preference change When the performance is lower than expected, we extract innovation themes and analyze and obtain the central values ​​of several innovation themes. , by the center value Screen out key themes from several innovation themes; Collect feedback data on the dissemination of innovative content containing key topics, and generate dissemination scores based on the dissemination feedback data , by the propagation Filter out target projects from the innovation content, classify the innovation paths of the generated target projects, obtain several path classes, and send push instructions to the outside; Push innovation support resources for each path type, record and obtain corresponding execution status data, and generate project completion from execution feedback data. If the project completion does not exceed the completion threshold, the corresponding innovation path will be used as a path to be adjusted; After collecting the feedback data of the target youth on path adjustment, the optimization features are extracted. The innovation path optimization knowledge graph outputs the optimization plan for the innovation path based on the optimization features. When the feasibility of the optimization plan is insufficient, innovation support resources are pushed for the optimized innovation path. The pre-trained K-Means algorithm extracts the interest preference data of target teenagers from the innovation activity data and marks them with timestamps; Sort several consecutive interest preferences along the time axis to obtain the corresponding preference column, use time series analysis to predict the interest preference, and then add the obtained prediction results to the preference column; Summarize the innovative content output by the target youth in the current stage, including text data and multimedia content data; after receiving the topic extraction instruction, use the pre-trained LDA algorithm to identify the innovative topics in the innovative content data, and use the trained deep learning model to add corresponding labels to the innovative content based on the extracted innovative topics; After arranging several innovation topics, a topic network diagram is constructed, and similarity analysis is performed between each innovation topic to obtain the corresponding similarity data; Calculate the central values ​​of several innovation topics through the obtained similarity data , if the center value Exceeding the central threshold, the innovation theme is considered as a key theme; the central value The generation method is as follows: ； Where: is the number of innovation themes, and is the weight coefficient, is the variance of similarity, weight coefficient: , ,and ; For the theme and The similarity between is the corresponding time interval, is the time decay function, is the covariance between similarity and time interval; When the innovative content containing key themes is in the dissemination state, the corresponding dissemination feedback data is collected and summarized as a dissemination feedback data set, and the dissemination degree is generated from the dissemination feedback data in the dissemination feedback data set. , if the propagation If the number is higher than the preset propagation threshold, it will be targeted as a project; After the target project determines the identity information of the corresponding target youth, collect the learning and innovation data of the target youth, including innovation content, learning content and knowledge and technology reserves, and generate an innovation learning data set after summarizing the current innovation resources and interest preference data; Taking the innovative learning data as input, the pre-trained path planning algorithm is used to plan the innovative path for the target youth, generating a personalized innovative path suitable for the target youth; After aggregating several personalized innovation paths, use a pre-trained classification algorithm to classify the personalized innovation paths, obtain several path classes, and send push instructions to the outside.

2. The method for managing youth innovation information according to claim 1, characterized in that: Collect various innovative contents uploaded by target teenagers, including video images and text content, use NLP algorithms to analyze the text content, and extract the interests, hobbies and emotional tendencies of target teenagers; Analyze the content themes, colors and character activities in video images, and summarize them to generate innovative activity data sets.

3. The method for managing youth innovation information according to claim 2, characterized in that: Calculate the preference change of interest preference based on continuous similarity data , if the preference change If the change degree is lower than the threshold, a topic extraction instruction is issued to the outside; the preference change degree The generation method is as follows: ； in, is the i-th similarity, is the mean of similarity, Standard deviation of similarity; weight coefficient: , ,and .

4. The method for managing youth innovation information according to claim 3 is characterized by: Under dimensionless conditions, the propagation degree is generated from the propagation feedback data The way is as follows: ； Where: , is the weight coefficient of time t and spatial position x, is the time-dependent weight coefficient, is the number of likes at time t; is the time attenuation coefficient, is the second-order rate of change of the forwarding amount in time and space, is the second-order rate of change of page views in time and space, For the propagation cycle.

5. The method for managing youth innovation information according to claim 4 is characterized by: Several innovation support resources are aggregated in advance to generate an innovation resource library. After receiving the push instruction, the innovation support resources are pushed to each path type based on the collaborative filtering algorithm to complete the push, collect push feedback data, and iteratively optimize the current resource push strategy based on the collected feedback data.

6. The method for managing youth innovation information according to claim 5, characterized in that: Record the execution status data of innovation projects, including the execution progress and target achievement rate of target projects; Taking the execution status data as input, the trained completion evaluation model is used to output the project completion. If the obtained completion does not exceed the completion threshold, the corresponding innovation path is taken as the path to be adjusted, and adjustment instructions are issued to the outside.

7. The method for managing youth innovation information according to claim 6 is characterized by: After receiving the adjustment instruction, after the path to be optimized is matched with the corresponding target youth, the target youth's innovation content dissemination data, resource support data, target project execution data and project feedback data are collected and summarized as path adjustment feedback data; Conduct principal component analysis on path adjustment feedback data and completion, extract key factors that affect the completion of target project execution, use youth innovation content optimization as the target word, and pre-construct an innovation path optimization knowledge graph.

8. The method for managing youth innovation information according to claim 7, characterized in that: Use the trained solution prediction model to predict the optimization solution, obtain the corresponding prediction data, and determine the expected completion degree of the optimized innovation path based on the prediction data; If the expected completion rate still does not exceed expectations, after executing the optimization plan to optimize the path to be adjusted, innovation support resources are pushed to the optimized innovation path based on the collaborative filtering algorithm.

Images