Service platform data management system and method based on artificial intelligence

Abstract

The invention discloses a service platform data management system and method based on artificial intelligence, and relates to the technical field of service platform data management, and the method comprises the steps: S10, analyzing the design quantitative feature data of each implementation design sub-idea; s20, extracting design learning interaction elements of the user; s30, analyzing association conditions between the extracted design learning interaction elements and the teaching videos uploaded to the education service platform and stored in the education service platform; and S40, the education service platform pushes the matched teaching video to a display interface of the user terminal. According to the method, the execution understanding fluctuation condition of the user on each design style and each design space feature is analyzed through the design quantitative feature data for uploading each implementation design sub-thought, the design learning interaction elements of the user are extracted in combination with the deviation condition of the user on each implementation design sub-thought, and the user experience is improved based on the extracted content. And weak spots which are not found by the user can be found.

Description

Technical Field

[0001] The present invention relates to the technical field of service platform data management, and specifically, to an artificial intelligence-based service platform data management system and method. Background Art

[0002] Educational service platforms have played an important role in aspects such as the integration and sharing of educational resources, teaching support, policy publicity and supervision, and international development, promoting the innovation and development of education.

[0003] Educational service platforms cover various types of teaching videos. For the selection of teaching videos, keyword retrieval is usually relied on to narrow the selection range. The number of teaching videos obtained by users through keyword retrieval is large but the accuracy is low. At the same time, potential weak points that users cannot detect still cannot be improved through teaching videos. Summary of the Invention

[0004] The purpose of the present invention is to provide an artificial intelligence-based service platform data management system and method to solve the problems raised in the prior art.

[0005] To achieve the above purpose, the present invention provides the following technical solution: An artificial intelligence-based service platform data management method, the method comprising: S10: Obtain the design idea basic data and design idea implementation data uploaded by the user to the educational service platform, and analyze the design quantification feature data of each implemented design sub-idea in combination with the implementation situation of the user for each design sub-idea; S20: Predict the learning index of the user for various design styles and design space features according to the design quantification feature data of each implemented design sub-idea, and extract the design learning interaction elements of the user based on the prediction results; S30: Analyze the association between the extracted design learning interaction elements and each teaching video stored in the educational service platform; S40: The educational service platform pushes the matching teaching videos to the user terminal display interface according to the analysis results.

[0006] Further, the S10 includes: S101: Obtain the design idea basic data and design idea implementation data uploaded by the user to the educational service platform; The design idea basic data includes the basic text description of each design sub-idea by the user, the text description of the design style of each design sub-idea by the user, and the text description of the spatial characteristics of each sub-design idea by the user; The data for implementing design ideas includes the implementation degree of each design sub-idea by the user, the written description of the design style corresponding to each implemented design sub-idea, and the written description of the spatial characteristics of each implemented design sub-idea. The implementation degree = the number of design features that have been implemented by the user in the design sub-idea / the total number of design features existing in the design sub-idea; S102: The data of design quantitative characteristics includes the innovation difficulty quantitative characteristics, satisfaction quantitative characteristics, and iteration quantitative characteristics of each implemented design sub-idea by the user; Number the basic written descriptions of each implemented sub-idea uploaded by the user to the education service platform. The numbering result is: i = 1, 2, …, n; n represents the total number of numbers. In the design database, obtain the number f of design works similar to the implemented design sub-idea with the number i uploaded by the user i and the average score g of the design works similar to the implemented design sub-idea with the number i by professional technical reviewers. i for acquisition; According to C i = L i *α + [1 - exp(-f i )]*β to calculate the innovation difficulty quantitative characteristic value C i of the implemented design sub-idea with the number i. Among them, L i represents the implementation degree of the implemented design sub-idea with the number i, α and β both represent weight coefficients and α + β = 1, exp() represents the exponential function with e as the base and e = 2.73; The satisfaction quantitative characteristic value M i of the implemented design sub-idea with the number i = 1 - exp(-g i ); The iteration quantitative characteristic value D i of the implemented design sub-idea with the number i = 1 - exp(-s i ), where s i represents the number of times the user has drawn the implemented design sub-idea with the number i. By analyzing the user's design ideas from the three aspects of the user's innovation difficulty, the external satisfaction with the design idea, and the user's iteration of the design idea, it is beneficial to analyze the user's understanding degree of the implementation of the design idea. Compared with directly analyzing the user's understanding of the implementation of the design idea based on the written description of the design idea or the written description or image similar to the image corresponding to the design idea, the design style of the user's design idea is retained.

[0007] Furthermore, the specific method for S20 to extract the design learning interaction elements of each implemented design sub-idea by the user is: According to the text descriptions of the design styles for implementing each design sub-idea and the text descriptions of their spatial characteristics, classify each implemented design sub-idea to obtain several classification subsets; number each classification subset, and the numbering result is: j = 1, 2, …, v; v represents the total number of obtained classification subsets; re-number the implemented design sub-ideas stored in each classification subset, and the numbering result is: p = 1, 2, …, q; q represents the total number of implemented design sub-ideas stored in the classification subset. For the innovation difficulty quantification eigenvalue C jp of the implemented design sub-idea p stored in classification subset j jp and the satisfaction quantification eigenvalue M jp and the iteration quantification eigenvalue D obtain them. jp According to the calculation method of the average value, calculate the average innovation difficulty quantification eigenvalue C´ jp corresponding to classification subset j jp and the average satisfaction quantification eigenvalue M´ j and the average iteration quantification eigenvalue D´ j respectively; according to the variation value = standard deviation / average value, calculate the innovation difficulty quantification eigenvalue variation c j corresponding to classification subset j and the satisfaction quantification eigenvalue variation m jp and the iteration quantification eigenvalue variation d j respectively. Taking the reciprocal of C´ jp as the base and c j as the exponent, calculate the innovation fluctuation index of the user for the design style and design space characteristics corresponding to classification subset j, and calculate the product with 1 - exp( - q j ) to obtain the innovation difficulty quantification eigenvalue learning coefficient c´ j corresponding to classification subset j. Similarly: According to the product with 1 - exp( - q j ), calculate the satisfaction quantification eigenvalue learning coefficient m´ j corresponding to classification subset j; according to the product with 1 - exp( - q j ), calculate the iteration quantification eigenvalue learning coefficient d´ j corresponding to classification subset j. Analyze the fluctuation of users' execution understanding of each design style and design space feature through the calculated feature change values and feature average values, and combine the users' preference for various design ideas to analyze the learning situation of users for each design style and design space feature. Compared with the educational videos directly screened by the educational service platform according to the keywords uploaded by users, it can not only accurately meet the users' needs, but also find the weak points that users have not noticed. According to K j =c´ j *u1+m´ j *u2+d´ j *u3, calculate the learning index of users for the design style and design space features corresponding to the classification subset j, and use the design style and design space features corresponding to the classification subset corresponding to maxK j as the design learning interaction elements of users. Among them, u1, u2, and u3 all represent influence coefficients and u1 + u2 + u3 = 1.

[0008] Further, the S30 includes: S301: Extract the key slice videos in the teaching videos stored in the educational service platform. The specific extraction method is as follows: Randomly select a teaching video in the educational service platform, mark the time data of the educational video in dimensions of theoretical explanation, visual case, innovation analysis, and comparative analysis. According to the time data of the educational video marked by the theoretical explanation dimension, obtain the first key slice video. According to the time data of the educational video marked by the visual case dimension, obtain the second key slice video. According to the time data of the educational video marked by the innovation analysis dimension, obtain the third key slice video. According to the time data of the educational video marked by the comparative analysis dimension, obtain the fourth key slice video; Traverse all the teaching videos in the educational service platform and extract the first key slice video, the second key slice video, the third key slice video, and the fourth key slice video in each teaching video; S302: Compare the definitions of the design style and design space features in the first key slice video of each teaching video with the extracted design learning interaction elements. If the comparison is successful, retain the corresponding teaching video. If the comparison is unsuccessful, do not retain the corresponding teaching video; Obtain the learning time periods of the second key slice video, the third key slice video, and the fourth key slice video in each retained teaching video for each historical user, and determine the concentrated learning time periods of the second key slice video, the third key slice video, and the fourth key slice video in each retained teaching video for historical users; S303: Number the remaining teaching videos, and the numbering result is: x = 1, 2, …, X; X represents the total number of remaining teaching videos. Calculate the ratio Y between the centralized learning time t of the historical user for the second key slice video in teaching video x and the learning time T corresponding to the second key slice video in teaching video x. Calculate according to F = m´ * Y + c´ * Y + d´ * Y. Calculate the correlation coefficient between the extracted design learning interaction elements and teaching video x. Among them, Y is the ratio between the centralized learning time t of the historical user for the third key slice video in teaching video x and the learning time T corresponding to the third key slice video in teaching video x, and Y is the ratio between the centralized learning time t of the historical user for the fourth key slice video in teaching video x and the learning time T corresponding to the fourth key slice video in teaching video x. By performing multi-dimensional marking on the teaching videos, obtaining several key slice videos, and combining the learning situations of historical users for each key slice video, analyze the correlation between the teaching videos and the extracted design learning interaction elements, perform classification comparison, and further improve the accuracy of finding matching teaching videos. 2x and the learning time T corresponding to the second key slice video in teaching video x 2x The ratio Y 2x is calculated. According to F x = m´ r * Y 2x + c´ r * Y 3x + d´ r * Y 4x Calculate the correlation coefficient between the extracted design learning interaction elements and teaching video x. Among them, Y 3x is the centralized learning time t of the historical user for the third key slice video in teaching video x 3x and the learning time T corresponding to the third key slice video in teaching video x 3x The ratio between them, Y 4x is the centralized learning time t of the historical user for the fourth key slice video in teaching video x 4x and the learning time T corresponding to the fourth key slice video in teaching video x 4x The ratio between them. By performing multi-dimensional marking on the teaching videos, obtaining several key slice videos, and combining the learning situations of historical users for each key slice video, analyze the correlation between the teaching videos and the extracted design learning interaction elements, perform classification comparison, and further improve the accuracy of finding matching teaching videos.

[0009] Furthermore, the S40 includes: Determine the number corresponding to maxF x and record the determined number as y, where y = 1, 2, …, X and y ≠ x. Then, teaching video y is called the matching teaching video for the user, and the education service platform pushes teaching video y to the user terminal display interface.

[0010] An artificial intelligence-based service platform data management system, the system includes a design data acquisition module, a design quantization feature analysis module, a learning coefficient prediction module, a design learning interaction element extraction module, a correlation analysis module, and an education service platform management module; The design data acquisition module is used to acquire the design idea basic data and design idea implementation data uploaded by the user to the education service platform; The design quantization feature analysis module analyzes the design quantization feature data of each implemented design sub-idea according to the implementation situation of the user for each design sub-idea; The design learning interaction element extraction module is used to extract the design learning interaction elements of the user according to the design quantitative feature data of each implemented design sub-idea; The correlation analysis module is used to analyze the correlation between the extracted design learning interaction elements and each teaching video uploaded and stored in the education service platform; The education service platform management module is used to push the matched teaching videos to the display interface of the user terminal.

[0011] Further, the design quantitative feature analysis module includes an innovation difficulty quantitative feature analysis unit, a satisfaction quantitative feature analysis unit, and an iteration quantitative feature analysis unit; The innovation difficulty quantitative feature analysis unit calculates the innovation difficulty quantitative feature values of each implemented design sub-idea according to the implementation degree of the user for each implemented design sub-idea and the number of design works similar to each implemented design sub-idea uploaded by the user in the design database; The satisfaction quantitative feature analysis unit calculates the satisfaction quantitative feature values of each implemented design sub-idea according to the average score of the professional technical review for the design works similar to each implemented design sub-idea; The iteration quantitative feature analysis unit calculates the iteration quantitative feature values of each implemented design sub-idea according to the number of times the user draws each implemented design sub-idea.

[0012] Further, the learning index prediction module includes a classification unit, a change value calculation unit, a learning coefficient prediction unit, and a design learning interaction element extraction unit; The classification unit classifies each implemented design sub-idea according to the design style text description and the feature text description in space of each implemented design sub-idea, and obtains several classification subsets; The change value calculation unit calculates the innovation difficulty quantitative feature change value, the satisfaction quantitative feature change value j and the iteration quantitative feature change value corresponding to each classification subset according to the change value = standard deviation / average value; The learning coefficient prediction unit predicts the innovation difficulty quantitative feature learning coefficient, the iteration quantitative feature learning coefficient, and the satisfaction quantitative feature learning coefficient of the design style and design space features corresponding to each classification subset according to the constructed mathematical model; The design learning interaction element extraction unit extracts the design learning interaction elements of the user according to the learning index of the design style and design space features corresponding to the classification subset j.

[0013] Further, the correlation analysis module includes a key slice video extraction unit, a teaching video screening and retention unit, and a correlation coefficient calculation unit; The key slice video extraction unit marks the time data of educational videos in dimensions of theoretical explanation, visual cases, innovation analysis, and comparative analysis. Based on the marking results, the first key slice video, the second key slice video, the third key slice video, and the fourth key slice video in each teaching video are extracted; The teaching video screening and retention unit compares the definitions of design styles and design space features in the first key slice video of each teaching video with the extracted design learning interaction elements. Based on the comparison results, the teaching videos stored in the educational service platform are screened and retained; The correlation coefficient calculation unit calculates the correlation coefficients between the extracted design learning interaction elements and each teaching video according to the constructed mathematical formula.

[0014] Compared with the prior art, the beneficial effects of the present invention are: 1. The present invention analyzes the execution understanding fluctuations of users for each design style and design space feature through the design quantitative feature data of each implemented design sub-idea for uploading, combines the preferences of users for various implemented design sub-ideas, extracts the design learning interaction elements of users. Based on the extracted content, it can also find the weak points that users have not noticed. In addition, using the extracted design learning interaction elements as keywords to preliminarily screen teaching videos increases the accuracy of keywords. Compared with the prior art, it improves the search accuracy of teaching videos.

[0015] 2. The present invention effectively identifies the teaching highlights in teaching videos by multi-dimensionally marking teaching videos and obtaining multiple key slice videos, and combines the learning coefficients of the respective quantitative features corresponding to the extracted design learning interaction elements to calculate the correlation coefficients between the extracted design learning interaction elements and each teaching video. Based on the calculation results, the matching teaching videos for users are determined, achieving accurate search for teaching videos, and the number of searched teaching videos is small, which is convenient for users to accurately learn.

[0016] 3. The present invention simplifies the analysis complexity of teaching video data by marking and dividing the time data of teaching videos, which is beneficial for the educational service platform to effectively manage teaching service data. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 It is a schematic work flow diagram of a data management method for an artificial intelligence-based service platform of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0018] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with 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 based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0019] As Figure 1 shown, the present invention provides a technical solution for a data management system and method of a service platform based on artificial intelligence. A method for managing data of a service platform based on artificial intelligence includes: S10: Obtain the basic data of the design idea and the implemented data of the design idea uploaded by the user to the education service platform, and analyze the design quantization feature data of each implemented design sub-idea in combination with the implementation situation of the user for each design sub-idea. S10 includes: S101: Obtain the basic data of the design idea and the implemented data of the design idea uploaded by the user to the education service platform. The basic data of the design idea includes the basic text description of each design sub-idea by the user, the text description of the design style of each design sub-idea by the user. The design style includes cubism style, abstractism style, and realism style, and the text description of the spatial characteristics of each sub-design idea by the user. The text description of the spatial characteristics of the design sub-idea includes plane characteristics and depth space characteristics. The plane characteristics refer to the visual characteristics and expression forms presented in the two-dimensional space (i.e., the plane), and the depth space characteristics refer to the spatial effect with depth and layering created by visual elements. The implemented data of the design idea includes the implementation degree of each design sub-idea by the user, the text description of the design style corresponding to each implemented design sub-idea, and the text description of the spatial characteristics of each implemented design sub-idea. The implemented design sub-idea refers to the design sub-idea with an implementation degree greater than 0. The implementation degree = the number of design features that have been implemented by the user in the design sub-idea / the total number of design features existing in the design sub-idea. The design features are determined according to the basic text description of each design sub-idea. For example: Suppose the text description of the design sub-idea A uploaded by the user to the education service platform is: Bury the part of the chair except the backrest and armrest in the snow and reshape the chair form through the coverage of the snow. The design features existing in the design sub-idea A are snow, chair backrest, and chair armrest. Suppose the design features snow and chair armrest have been implemented by the user, then the implementation degree of the user for the design sub-idea A = 2 / 3 = 0.67. S102: The design quantization feature data includes the innovation difficulty quantization feature, satisfaction quantization feature, and iteration quantization feature of each implemented design sub-idea by the user. Number the basic text descriptions of each implementation sub-idea uploaded by the user to the education service platform. The numbering result is: i = 1, 2, …, n; n represents the total number of numbers. In the design database, obtain the number f of design works similar to the implementation design sub-idea numbered i uploaded by the user i , and the average score g of the design works similar to the implementation design sub-idea numbered i by professional technical reviewers i for acquisition; According to C i = L i *α + [1 - exp( - f i )]*β to calculate the innovation difficulty quantization eigenvalue C i of the implementation design sub-idea numbered i, where L i represents the implementation degree of the implementation design sub-idea numbered i, α and β both represent weight coefficients and α + β = 1, exp() represents the exponential function with base e and e = 2.73; The satisfaction quantization eigenvalue M i of the implementation design sub-idea numbered i i = 1 - exp( - g The iteration quantization eigenvalue D i of the implementation design sub-idea numbered i i = 1 - exp( - s i ), where s represents the number of times the user draws the implementation design sub-idea numbered i; S20: According to the design quantization characteristic data of each implementation design sub-idea, predict the learning index of the user for various design styles and design space characteristics. Based on the prediction results, extract the design learning interaction elements of the user; The specific method for S20 to extract the design learning interaction elements of the user for each implementation design sub-idea is as follows: According to the design style text description and the spatial characteristic text description of each implementation design sub-idea, classify each implementation design sub-idea and obtain several classification subsets. For the classification subsets with multiple implementation design sub-ideas, the design style text description and the spatial characteristic text description stored in the classification subsets are the same; number each classification subset. The numbering result is: j = 1, 2, …, v; v represents the total number of obtained classification subsets; re-number the implementation design sub-ideas stored in the classification subsets. The numbering result is: p = 1, 2, …, q; q represents the total number of implementation design sub-ideas stored in the classification subset; For the innovation difficulty quantization eigenvalue C jp , satisfaction quantization eigenvalue M jp and iteration quantization eigenvalue Djp Obtain; According to the calculation method of the average value, calculate the average value C´ of the innovation difficulty quantization feature corresponding to the classification subset j respectively jp , the average value M´ of the satisfaction quantization feature jp and the average value D´ of the iteration quantization feature jp The calculation method of the average value is the sum of a set of data divided by the number of data. For example: M´ jp =(M´ j1 +M´ j2 +…+M´ jq ) / q; According to the abnormal change value = standard deviation / average value, calculate the innovation difficulty quantization feature abnormal change value c j , the satisfaction quantization feature abnormal change value m j and the iteration quantization feature abnormal change value d j respectively. The satisfaction quantization feature abnormal change value refers to the fluctuation degree of the satisfaction quantization feature value data set corresponding to the implemented design sub-idea stored in the classification subset compared with the average value of the satisfaction quantization feature. The calculation method of the standard deviation belongs to the prior art; Taking the reciprocal of C´ jp as the base and c j as the exponent, calculate the innovation fluctuation index of the user for the design style and design space characteristics corresponding to the classification subset j, and calculate the product with 1-exp(-q j ), to obtain the innovation difficulty quantization feature learning coefficient c´ j corresponding to the classification subset j; Similarly: According to the product with 1-exp(-q j ), calculate the satisfaction quantization feature learning coefficient m´ j corresponding to the classification subset j; According to the product with 1-exp(-q j ), calculate the iteration quantization feature learning coefficient d´ j corresponding to the classification subset j; Among them, 1-exp(-q j ) represents the preference index of the user for the design style and design space characteristics corresponding to the classification subset j; represents the satisfaction fluctuation index of the professional technical review for the design style and design space characteristics implemented by the user corresponding to the classification subset j, represents the iteration fluctuation index of the user for the design style and design space characteristics corresponding to the classification subset j; According to K j =c´ j *u1+m´ j*u2 + d´ j *u3 calculates the learning index of the user for the design style and design space features corresponding to the classification subset j, and takes maxK j The design style and design space features corresponding to the corresponding classification subset are used as the user's design learning interaction elements, where u1, u2, and u3 all represent influence coefficients and u1 + u2 + u3 = 1, and K j represents the learning index of the user for the design style and design space features corresponding to the classification subset j. The design space features include planar features and depth space features; S30: Analyze the association between the extracted design learning interaction elements and each teaching video stored in the educational service platform; S30 includes: S301: Extract the key slice videos in the teaching videos stored in the educational service platform. The specific extraction method is as follows: Randomly select a teaching video in the educational service platform, and mark the educational video time data with dimensions of theoretical explanation (used to describe the definition and historical background of design style and design space features), visual case (used to disassemble typical works), innovation analysis (used to analyze the innovative designs in typical works), and comparative analysis (used to compare and analyze various design concepts of different design styles and design space features). According to the educational video time data marked by the theoretical explanation dimension, obtain the first key slice video. According to the educational video time data marked by the visual case dimension, obtain the second key slice video. According to the educational video time data marked by the innovation analysis dimension, obtain the third key slice video. According to the educational video time data marked by the comparative analysis dimension, obtain the fourth key slice video. For example, if the educational video time data marked by the theoretical explanation dimension are 12:00, 15:00, 16:00, and 19:00, then the first key slice video refers to the slice video selected from the teaching video with the video playback time between 12:00 and 15:00 and the slice video selected from the teaching video with the video playback time between 16:00 and 19:00. The learning time corresponding to the first key slice video in the teaching video is 19:00 - 16:00 + 15:00 - 12:00 = 6 minutes; Traverse all the teaching videos in the educational service platform, and extract the first key slice video, the second key slice video, the third key slice video, and the fourth key slice video in each teaching video; S302: Compare the definitions of the design style and design space features in the first key slice video of each teaching video with the extracted design learning interaction elements. If the comparison is successful, retain the corresponding teaching video; if the comparison is unsuccessful, do not retain the corresponding teaching video. A successful comparison means that the definitions of the design style and design space features in the first key slice video of the teaching video are the same as the extracted definitions of the design style and design space features. Obtain the learning time periods of the second key slice video, the third key slice video, and the fourth key slice video in each retained teaching video for each historical user, and determine the concentrated learning time periods of the historical users for the second key slice video, the third key slice video, and the fourth key slice video in each retained teaching video. The concentrated learning time periods are obtained by taking the intersection of the learning time periods. S303: Number the retained teaching videos, and the numbering result is: x = 1, 2, …, X; X represents the total number of retained teaching videos. Calculate the ratio Y between the concentrated learning time t of the historical user for the second key slice video in teaching video x and the learning time T corresponding to the second key slice video in teaching video x. Calculate the correlation coefficient between the extracted design learning interaction elements and teaching video x according to F = m´ * Y + c´ * Y + d´ * Y. Here, Y is the ratio between the concentrated learning time t of the historical user for the third key slice video in teaching video x and the learning time T corresponding to the third key slice video in teaching video x, and Y is the ratio between the concentrated learning time t of the historical user for the fourth key slice video in teaching video x and the learning time T corresponding to the fourth key slice video in teaching video x. m´, c´, and d´ respectively represent the satisfaction quantization feature learning coefficient, innovation difficulty quantization feature learning coefficient, and iteration quantization feature learning coefficient corresponding to the classification subset corresponding to maxK. 2x And the learning time T corresponding to the second key slice video in teaching video x 2x The ratio Y between them 2x Calculate according to F x = m´ r * Y 2x + c´ r * Y 3x + d´ r * Y 4x Calculate the correlation coefficient between the extracted design learning interaction elements and teaching video x. Here, Y 3x Is the ratio between the concentrated learning time t of the historical user for the third key slice video in teaching video x and the learning time T corresponding to the third key slice video in teaching video x 3x And the learning time T corresponding to the third key slice video in teaching video x 3x The ratio between them, Y 4x Is the ratio between the concentrated learning time t of the historical user for the fourth key slice video in teaching video x and the learning time T corresponding to the fourth key slice video in teaching video x 4x And the learning time T corresponding to the fourth key slice video in teaching video x 4x The ratio between them, m´ r 、c´ r 、d´ r Respectively represent the satisfaction quantization feature learning coefficient, innovation difficulty quantization feature learning coefficient, and iteration quantization feature learning coefficient corresponding to the classification subset corresponding to maxK j ; S40: The education service platform pushes the matched teaching videos to the user terminal display interface according to the analysis results. S40 includes: For maxFx Determine the corresponding number, and denote the determined number as y, where y = 1, 2, …, X and y ≠ x. Then, the teaching video y is called the matching teaching video for the user, and the education service platform pushes the teaching video y to the display interface of the user terminal.

[0020] A data management system for a service platform based on artificial intelligence. The system includes a design data acquisition module, a design quantization feature analysis module, a learning coefficient prediction module, a design learning interaction element extraction module, an association analysis module, and an education service platform management module; The design data acquisition module is used to acquire the basic design idea data and the implemented design idea data uploaded by the user to the education service platform; The design quantization feature analysis module analyzes the design quantization feature data of each implemented design sub-idea according to the implementation situation of the user for each design sub-idea; The design quantization feature analysis module includes an innovation difficulty quantization feature analysis unit, a satisfaction quantization feature analysis unit, and an iteration quantization feature analysis unit; The innovation difficulty quantization feature analysis unit calculates the innovation difficulty quantization feature values of each implemented design sub-idea according to the implementation degree of the user for each implemented design sub-idea and the number of design works similar to each implemented design sub-idea uploaded by the user in the design database; The satisfaction quantization feature analysis unit calculates the satisfaction quantization feature values of each implemented design sub-idea according to the average score of the professional technical review for the design works similar to each implemented design sub-idea; The iteration quantization feature analysis unit calculates the iteration quantization feature values of each implemented design sub-idea according to the number of drawing times of the user for each implemented design sub-idea; The design learning interaction element extraction module is used to extract the design learning interaction elements of the user according to the design quantization feature data of each implemented design sub-idea; The learning index prediction module includes a classification unit, a deviation value calculation unit, a learning coefficient prediction unit, and a design learning interaction element extraction unit; The classification unit classifies each implemented design sub-idea according to the text description of the design style and the text description of the features in space of each implemented design sub-idea, and obtains several classification subsets; The deviation value calculation unit calculates the innovation difficulty quantization feature deviation value, the satisfaction quantization feature deviation value j and the iteration quantization feature deviation value corresponding to each classification subset according to the deviation value = standard deviation / average value; The learning coefficient prediction unit predicts the innovation difficulty quantization feature learning coefficient, the iteration quantization feature learning coefficient, and the satisfaction quantization feature learning coefficient of the design style and design space features corresponding to each classification subset according to the constructed mathematical model; The design learning interaction element extraction unit extracts the user's design learning interaction elements according to the learning index of the design style and design space characteristics corresponding to the classification subset j by the user; The correlation analysis module is used to analyze the correlation between the extracted design learning interaction elements and each teaching video uploaded and stored in the education service platform; The correlation analysis module includes a key slice video extraction unit, a teaching video screening and retention unit, and a correlation coefficient calculation unit; The key slice video extraction unit marks the time data of the education video in dimensions of theoretical explanation, visual case, innovation analysis, and comparative analysis. Based on the marking results, the first key slice video, the second key slice video, the third key slice video, and the fourth key slice video in each teaching video are extracted; The teaching video screening and retention unit compares the definitions of the design style and design space characteristics in the first key slice video of each teaching video with the extracted design learning interaction elements. Based on the comparison results, the teaching videos stored in the education service platform are screened and retained; The correlation coefficient calculation unit calculates the correlation coefficient between the extracted design learning interaction elements and each teaching video according to the constructed mathematical formula; The education service platform management module is used to push the matching teaching videos to the display interface of the user terminal.

[0021] For those skilled in the art, it is obvious that the present invention is not limited to the details of the above exemplary embodiments, and without departing from the spirit or basic characteristics of the present invention, the present invention can be implemented in other specific forms. Therefore, from any point of view, the embodiments should be regarded as exemplary and non-limiting. The scope of the present invention is defined by the appended claims rather than the above description. Therefore, all changes falling within the meaning and scope of the equivalent elements of the claims are intended to be included in the present invention. Any reference signs in the claims should not be regarded as limiting the claimed rights.

Claims

1. A method for managing service platform data based on artificial intelligence, characterized in that: The method includes: S10: Obtain the basic data of the design idea and the data of the implementation of the design idea uploaded by the user to the education service platform, and analyze the design quantitative feature data of each implemented design sub-idea in combination with the implementation situation of the user for each design sub-idea; S20: Predict the learning index of the user for various design styles and design space features according to the design quantitative feature data of each implemented design sub-idea, and extract the design learning interaction elements of the user based on the prediction results; S30: Analyze the association between the extracted design learning interaction elements and each teaching video stored in the education service platform; S40: The education service platform pushes the matching teaching videos to the display interface of the user terminal according to the analysis results.

2. The data management method of a service platform based on artificial intelligence according to claim 1, wherein: The S10 includes: S101: Obtain the basic data of the design idea and the data of the implementation of the design idea uploaded by the user to the education service platform; The basic data of the design idea includes the basic text description of each design sub-idea by the user, the text description of the design style of each design sub-idea by the user, and the text description of the spatial characteristics of each sub-design idea by the user; The data of the implementation of the design idea includes the implementation degree of each design sub-idea by the user, the text description of the design style corresponding to each implemented design sub-idea, and the text description of the spatial characteristics of each implemented design sub-idea. The implementation degree = the number of design features that have been implemented by the user in the design sub-idea / the total number of design features existing in the design sub-idea; S102: The design quantitative feature data includes the innovation difficulty quantitative feature, the satisfaction quantitative feature, and the iteration quantitative feature of each implemented design sub-idea by the user; Number the basic text descriptions of each implemented sub-idea uploaded by the user to the education service platform. The numbering results are: i = 1, 2, …, n; n represents the total number of numbers. In the design database, obtain the number f of design works similar to the implemented design sub-idea numbered i uploaded by the user i , and the average score g of the design works similar to the implemented design sub-idea numbered i by professional technical reviewers i for acquisition; According to C i =L i *α + [1 - exp(-f i )]*β to calculate the innovation difficulty quantification eigenvalue C i for the implementation design sub-idea numbered i, where L i represents the implementation degree of the implementation design sub-idea numbered i, α and β both represent weight coefficients and α + β = 1, exp() represents the exponential function with e as the base and e = 2.73; Satisfaction quantification eigenvalue M of the implementation design sub-idea numbered i i = 1 - exp(-g i ); The iterative quantization eigenvalue D for implementing the design sub-idea numbered i i = 1 - exp(-s i ), where s i represents the number of times the user has drawn the design sub-idea numbered i.

3. A data management method for an artificial intelligence-based service platform according to claim 2, characterized in that: The specific method for the S20 to extract the design learning interaction elements of each implemented design sub-idea by the user is: Classify each implemented design sub-idea according to the text description of the design style and the text description of the spatial characteristics of each implemented design sub-idea, and obtain several classification subsets; number each classification subset, and the numbering result is: j = 1, 2,..., v; v represents the total number of obtained classification subsets; re-number each implemented design sub-idea stored in the classification subset, and the numbering result is: p = 1, 2,..., q; q represents the total number of implemented design sub-ideas stored in the classification subset; Obtain the innovation difficulty quantification eigenvalue C for implementing the design sub-idea p stored in the classification subset j jp , the satisfaction quantification eigenvalue M jp and the iteration quantification eigenvalue D jp ; Calculate the average value of the innovation difficulty quantification feature C´ corresponding to the classification subset j, the average value of the satisfaction quantification feature M´ jp , and the average value of the iteration quantification feature D´ jp respectively according to the calculation method of the average value; calculate the anomaly value of the innovation difficulty quantification feature c jp , the anomaly value of the satisfaction quantification feature m j , and the anomaly value of the iteration quantification feature d j corresponding to the classification subset j respectively according to the anomaly value = standard deviation / average value; j ​ With the reciprocal of C´ jp as the base and c j as the exponent, calculate the innovation fluctuation index of the user's design style and design space characteristics corresponding to the classification subset j. Calculate the product between and 1 - exp(-q j ), and obtain the learning coefficient c´ of the innovation difficulty quantization feature corresponding to the classification subset j j ; Similarly, according to the product between 1 - exp(-q j ), the satisfaction quantization feature learning coefficient m' corresponding to the classification subset j j is calculated; according to the product between 1 - exp(-q j ), the iterative quantization feature learning coefficient d' corresponding to the classification subset j j is calculated; According to K j = c´ j * u1 + m´ j * u2 + d´ j * u3 to calculate the learning index of the user for the design style and design space features corresponding to the classification subset j, and take the design style and design space features corresponding to the classification subset corresponding to maxK j as the user's design learning interaction elements, where u1, u2, and u3 all represent influence coefficients and u1 + u2 + u3 = 1.

4. A data management method for an artificial intelligence-based service platform according to claim 3, characterized in that: The S30 includes: S301: Extract the key slice videos in the teaching videos stored in the education service platform. The specific extraction method is: randomly select a teaching video in the education service platform, mark the time data of the education video in dimensions of theoretical explanation, visual case, innovation analysis, and comparative analysis, obtain the first key slice video according to the time data of the education video marked by the theoretical explanation dimension, obtain the second key slice video according to the time data of the education video marked by the visual case dimension, obtain the third key slice video according to the time data of the education video marked by the innovation analysis dimension, and obtain the fourth key slice video according to the time data of the education video marked by the comparative analysis dimension; Traverse all teaching videos in the education service platform, and extract the first key slice video, the second key slice video, the third key slice video, and the fourth key slice video in each teaching video; S302: Compare the definitions of the design style and design space features in the first key slice video of each teaching video with the extracted design learning interaction elements. If the comparison is successful, retain the corresponding teaching video; if the comparison is unsuccessful, do not retain the corresponding teaching video; Obtain the learning time periods of the second key slice video, the third key slice video, and the fourth key slice video in each retained teaching video for each historical user, and determine the concentrated learning time periods of the second key slice video, the third key slice video, and the fourth key slice video in each retained teaching video for the historical user; S303: Number the remaining teaching videos, and the numbering result is: x = 1, 2, …, X; X represents the total number of remaining teaching videos. Calculate the ratio Y between the concentrated learning time t of the historical users for the second key slice video in the teaching video x and the learning time T corresponding to the second key slice video in the teaching video x. Calculate according to F = m´ * Y + c´ * Y + d´ * Y. Calculate the correlation coefficient between the extracted design learning interaction elements and the teaching video x. Among them, Y is the ratio between the concentrated learning time t of the historical users for the third key slice video in the teaching video x and the learning time T corresponding to the third key slice video in the teaching video x, and Y is the ratio between the concentrated learning time t of the historical users for the fourth key slice video in the teaching video x and the learning time T corresponding to the fourth key slice video in the teaching video x. 2x and the learning time T corresponding to the second key slice video in the teaching video x 2x The ratio Y 2x is calculated. According to F x = m´ r * Y 2x + c´ r * Y 3x + d´ r * Y 4x Calculate the correlation coefficient between the extracted design learning interaction elements and the teaching video x. Among them, Y 3x is the concentrated learning time t of the historical users for the third key slice video in the teaching video x 3x and the learning time T corresponding to the third key slice video in the teaching video x 3x The ratio, Y 4x is the concentrated learning time t of the historical users for the fourth key slice video in the teaching video x 4x and the learning time T corresponding to the fourth key slice video in the teaching video x 4x The ratio.

5. A data management method for an artificial intelligence-based service platform according to claim 4, characterized in that: The S40 includes: Determine the corresponding number for maxF x Determine the corresponding number, denote the determined number as y, where y = 1, 2, …, X and y ≠ x. Then, the teaching video y is called the matching teaching video for the user, and the educational service platform will push the teaching video y to the display interface of the user terminal.

6. An artificial intelligence-based service platform data management system applied to the artificial intelligence-based service platform data management method according to any one of claims 1-5, characterized in that: The system includes a design data acquisition module, a design quantitative feature analysis module, a learning coefficient prediction module, a design learning interaction element extraction module, a correlation analysis module, and an education service platform management module; The design data acquisition module is used to acquire the design idea basic data and design idea implementation data uploaded by the user to the education service platform; The design quantitative feature analysis module analyzes the design quantitative feature data of each implemented design sub-idea according to the implementation situation of the user for each design sub-idea; The design learning interaction element extraction module is used to extract the design learning interaction elements of the user according to the design quantitative feature data of each implemented design sub-idea; The correlation analysis module is used to analyze the correlation between the extracted design learning interaction elements and each teaching video stored in the education service platform; The education service platform management module is used to push the matching teaching videos to the user terminal display interface.

7. The data management system of the artificial intelligence-based service platform according to claim 6, characterized in that: The design quantitative feature analysis module includes an innovation difficulty quantitative feature analysis unit, a satisfaction quantitative feature analysis unit, and an iteration quantitative feature analysis unit; The innovation difficulty quantitative feature analysis unit calculates the innovation difficulty quantitative feature values of each implemented design sub-idea according to the implementation degree of the user for each implemented design sub-idea and the number of design works similar to each implemented design sub-idea uploaded by the user in the design database; The satisfaction quantitative feature analysis unit calculates the satisfaction quantitative feature values of each implemented design sub-idea according to the average score of the professional technical review for the design works similar to each implemented design sub-idea; The iteration quantitative feature analysis unit calculates the iteration quantitative feature values of each implemented design sub-idea according to the number of drawing times of the user for each implemented design sub-idea.

8. The data management system of the artificial intelligence-based service platform according to claim 7, characterized in that: The learning index prediction module includes a classification unit, a change value calculation unit, a learning coefficient prediction unit, and a design learning interaction element extraction unit; The classification unit classifies each implemented design sub-idea according to the design style text description and the spatial feature text description of each implemented design sub-idea, and obtains several classification subsets; The abnormal value calculation unit calculates the abnormal value of the innovation difficulty quantification feature, the satisfaction quantification feature j and the iterative quantification feature abnormal value corresponding to each classification subset according to the abnormal value = standard deviation / average value; The learning coefficient prediction unit predicts the innovation difficulty quantification feature learning coefficient, iterative quantification feature learning coefficient, and satisfaction quantification feature learning coefficient of the design style and design space features corresponding to each classification subset according to the constructed mathematical model; The design learning interaction element extraction unit extracts the design learning interaction elements of the user according to the learning index of the design style and design space features corresponding to the classification subset j by the user.

9. The data management system of an artificial intelligence-based service platform according to claim 8, characterized in that: The correlation analysis module includes a key slice video extraction unit, a teaching video screening and retention unit, and a correlation coefficient calculation unit; The key slice video extraction unit marks the time data of the educational video in dimensions of theoretical explanation, visual case, innovation analysis, and comparative analysis. Based on the marking results, the first key slice video, the second key slice video, the third key slice video, and the fourth key slice video in each teaching video are extracted; The teaching video screening and retention unit compares the definitions of the design style and design space features in the first key slice video of each teaching video with the extracted design learning interaction elements. Based on the comparison results, the teaching videos stored in the educational service platform are screened and retained; The correlation coefficient calculation unit calculates the correlation coefficient between the extracted design learning interaction elements and each teaching video according to the constructed mathematical formula.

Images