A cloud computing platform for educational service resources

Abstract

The present invention discloses an educational service resource cloud computing platform, which relates to the field of educational data processing technology, and includes an information acquisition module, a data processing module, a video evaluation module, a video recommendation module for calculating the recommendation score of candidate video resources in a video library, and a visual display module for recommending and displaying the candidate video resources with the highest recommendation score in the video library to learners. By providing a video recommendation module, the present invention can evaluate the adaptation value of each video resource in historical browsing data and current input data and the score value of the candidate video resource, calculate the recommendation score of the candidate video resource in the video library, so that when the platform makes recommendations to learners, the learner's historical browsing data can also be taken into account, thereby improving the accuracy of recommending videos to learners.

Description

Technical Field

[0001] The present invention relates to the technical field of educational data processing, and in particular to an educational service resource cloud computing platform. Background Art

[0002] With the rapid development of the Internet, we have entered the era of information explosion. Currently, there are more and more individuals or groups producing educational resources through the Internet. The types of resources provided (micro-courses, study plans, teaching plans, test questions, test papers, etc.) are endless, and the number of options is increasing. How to let people who need so many resources find them and meet the various needs of users is a difficult problem facing people.

[0003] However, in the prior art, the recommendation of educational resources is often achieved through simple screening, and the current recommendation of educational resources is based on the analysis of the learner's current input data and then recommended, without reference analysis of the learner's historical browsing data. This makes the recommended video resources for learners not accurate enough, which is not conducive to improving the reliability and convenience of video resource recommendations. Summary of the invention

[0004] In view of the deficiencies in the prior art, the purpose of the present invention is to provide an educational service resource cloud computing platform to solve the above-mentioned technical problems.

[0005] In order to achieve the above object, the present invention adopts the following technical solutions:

[0006] An educational service resource cloud computing platform, comprising:

[0007] The information acquisition module is used to obtain the learner's current input data and the learner's historical browsing data, and pre-process the historical browsing data, and then send the current input data and historical browsing data to the server for storage;

[0008] A data processing module, used for processing and calculating the current input data and the historical browsing data to obtain an adaptation value between each video resource in the historical browsing data and the current input data;

[0009] The video evaluation module is used to evaluate the candidate video resources in the video library to obtain the score value of the candidate video resources, specifically:

[0010] Preprocess the candidate video resources in the video library and obtain feature vectors of the candidate video resources;

[0011] Calculate the similarity value between the feature vector of the candidate video resource and the required feature vector, and screen the similarity values ​​to obtain a second similarity value;

[0012] The score of the candidate video resource is calculated by formula (III), which is as follows:

[0013]

[0014] where Df k Indicates the score value of the candidate video resource corresponding to the second similarity value. Represents the second similarity value of the selected candidate video resources, Cr k Represents the click rate of the selected candidate video resources, Sh k Pj represents the number of shares of the selected candidate video resources. k represents the feedback score value of the screened candidate video resources, ξ1 represents the similarity weight between the candidate video resources and the current input data, ξ2 represents the click rate weight of the candidate video resources, ξ3 represents the sharing weight of the candidate video resources, and ξ4 represents the feedback score weight of the candidate video, satisfying ξ1+ξ2+ξ3+ξ4=1, and ξ1 is greater than any one of ξ2, ξ3 and ξ4;

[0015] The similarity value between the feature vector of the candidate video resource and the required feature vector is calculated, and the similarity value is screened to obtain the second similarity value of the same type, which is specifically:

[0016] The similarity value between the feature vector of the candidate video resource and the required feature vector is calculated by formula (IV), and the formula is as follows:

[0017]

[0018] where cosα k Represents the cosine value between the feature vector of the candidate video resource and the required feature vector, α k represents the angle between the feature vector of the candidate video resource and the required feature vector, A represents the required feature vector, H k A feature vector representing a candidate video resource;

[0019] Compare the similarity value of the candidate video resource with a preset second similarity threshold interval, and if the similarity value belongs to the second similarity threshold interval, mark the similarity value as a second similarity value of the same type; if the similarity value does not belong to the second similarity threshold interval, discard the similarity value and the candidate video resource corresponding to the similarity value;

[0020] The video recommendation module is used to evaluate the adaptation value of each video resource in the historical browsing data and the current input data and the score value of the candidate video resource, and calculate the recommendation score of the candidate video resource in the video library;

[0021] The visualization display module is used to recommend and display the candidate video resources with the highest recommendation scores in the video library to learners.

[0022] Preferably, the information acquisition module performs the following steps to preprocess the historical browsing data:

[0023] Performing data cleaning on the historical browsing data to obtain a first data set;

[0024] Performing data conversion on the first data set to obtain a second data set;

[0025] Performing feature extraction on the second data set to obtain a third data set;

[0026] Perform data integration on the third data set to obtain a fourth data set.

[0027] Preferably, the data processing module processes and calculates the current input data and the historical browsing data to obtain the adaptation value of each video resource in the historical browsing data and the current input data in the following specific steps:

[0028] Obtain the search command currently input by the learner;

[0029] Using a text recognition algorithm to extract features from the search instruction to obtain a demand feature vector;

[0030] Obtaining video feature vectors corresponding to all video resources in the fourth data set;

[0031] Calculate the similarity value between each video resource feature vector and the demand feature vector, and filter the similarity values ​​to obtain a first similarity value of the same type;

[0032] Sum the similarity values ​​of the first category to obtain a total similarity value;

[0033] The adaptation value of the video resource is calculated by formula (1):

[0034]

[0035] in represents the adaptation value of the video resource, n represents the total number of similarity values ​​of the first type, and n is a positive integer. Indicates the first similarity value of the video feature vector.

[0036] Preferably, the similarity values ​​between each video feature vector and the requirement feature vector are calculated, and the similarity values ​​are screened to obtain a first similarity value of the same type, specifically:

[0037] The similarity value between each video resource feature vector and the demand feature vector is calculated by formula (2), and the formula is as follows:

[0038]

[0039] where cosθ i represents the cosine value between each video feature vector and the required feature vector, θ i represents the angle between each video feature vector and the required feature vector, A represents the required feature vector, B i represents the video feature vector;

[0040] The similarity value is compared with a preset first similarity threshold interval. If the similarity value belongs to the first similarity threshold interval, the similarity value is marked as a first similarity value of the same type; if the similarity value does not belong to the first similarity threshold interval, the similarity value and the video resource corresponding to the similarity value are discarded.

[0041] Preferably, the video recommendation module evaluates the adaptation value of each video resource in the historical browsing data and the current input data and the score value of the candidate video resource, and calculates the recommendation score of the candidate video resource in the video library, as follows:

[0042] Obtain the feature vector B of the video resource corresponding to the first similarity value i ;

[0043] Sort the second similarity values ​​in order of size, and extract the feature vectors H of the candidate video resources corresponding to the top n second similarity values ​​in order of absolute value from large to small. i ;

[0044] Calculate the eigenvector H i With the eigenvector B i and marked it as indirect similarity;

[0045] Get the eigenvector H i The indirect adaptation value of the corresponding candidate video resource;

[0046] The recommendation score of the candidate video resources in the video library is calculated by formula (V), which is as follows:

[0047]

[0048] where Df i Indicates the score value of the candidate video resources corresponding to the top n second similarity values. It represents the indirect adaptation value of the candidate video resources corresponding to the top n second similarity values, β1 represents the score weight, β2 represents the indirect adaptation weight, and β1+β2=1, β1 is greater than 1.5β2.

[0049] Preferably, the feature vector H is calculated i With the eigenvector B iThe similarity is marked as indirect similarity and is implemented by formula (six), which is as follows:

[0050]

[0051] where cosγ i Indicates the indirect similarity between the feature vector of the candidate video resource corresponding to the second similarity value of the top n and the video feature vector, γ i Indicates the angle between the feature vector of the candidate video resource corresponding to the top n second similarity values ​​and the video feature vector.

[0052] Preferably, the feature vector H is obtained i The indirect adaptation value of the corresponding candidate video resource is realized by formula (VII), which is as follows:

[0053]

[0054] Where n represents the total number of first similarity values.

[0055] Preferably, the visualization display module recommends and displays the candidate video resource with the highest recommendation score in the video library to the learner as follows:

[0056] Sort the recommendation scores of the candidate video resources corresponding to the top n second-class similarity values ​​in the video library;

[0057] Obtain the candidate video resource corresponding to the largest recommendation score, and recommend the candidate video resource to the learner;

[0058] Track learners' feedback scores, sharing times and click-through rates of the candidate video resources in real time and update the feedback score weight, sharing weight and click-through rate weight to achieve subsequent accurate recommendations for candidate video resources.

[0059] In summary, due to the adoption of the above technical solution, the beneficial effects of the present invention are:

[0060] The present invention can perform feature extraction on the learner's current input data, and extract historical video resources similar to the current input data by comparing the current input data with the historical browsing data, and then perform similarity analysis on the candidate videos in the video library with the learner's current input data and the historical video data, so that when recommending videos to the learner, the learner's past video resources can be considered for recommendation. The present invention is provided with a visual display module, so that the present invention can display the candidate video with the highest recommendation score to the learner, and can also track the feedback score, sharing number and click-through rate of the recommended candidate video in real time and update the feedback score weight, sharing weight and click-through rate weight, so as to realize the subsequent accurate recommendation of the candidate video resources. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0062] Figure 1 A composition diagram of an educational service resource cloud computing platform is shown.

[0063] Figure 2 A flow chart for calculating the adaptation value is shown.

[0064] Figure 3 A flow chart for calculating recommendation scores is shown.

[0065] Legend:

[0066] 1. Information acquisition module; 2. Data processing module; 3. Video evaluation module; 4. Video recommendation module; 5. Visual display module. DETAILED DESCRIPTION

[0067] 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.

[0068] In the description of the present invention, it should be understood that the terms "length", "width", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside" and "outside" etc., indicating orientations or positional relationships, are based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be understood as a limitation on the present invention.

[0069] It should be noted that when a component is referred to as being "fixed to" another component, it may be directly on the other component or there may also be a component centered. When a component is considered to be "connected to" another component, it may be directly connected to the other component or there may also be a component centered. When a component is considered to be "set on" another component, it may be directly set on the other component or there may also be a component centered. The terms "vertical", "horizontal", "left", "right" and similar expressions used herein are for illustrative purposes only.

[0070] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.

[0071] Reference Figure 1-Figure 3 , an embodiment of an educational service resource cloud computing platform of the present invention is further described.

[0072] An educational service resource cloud computing platform, comprising:

[0073] Information acquisition module 1, used to acquire the learner's current input data and the learner's historical browsing data, and pre-process the historical browsing data, and then send the current input data and historical browsing data to the server for storage;

[0074] Data processing module 2, used for processing and calculating the current input data and the historical browsing data to obtain the adaptation value of each video resource in the historical browsing data and the current input data;

[0075] The video evaluation module 3 is used to evaluate the candidate video resources in the video library to obtain the score value of the candidate video resources;

[0076] The video recommendation module 4 is used to evaluate the adaptation value of each video resource in the historical browsing data and the current input data and the score value of the candidate video resource, and calculate the recommendation score of the candidate video resource in the video library;

[0077] The visualization display module 5 is used to recommend and display the candidate video resources with the highest recommendation scores in the video library to the learners.

[0078] The steps of preprocessing the historical browsing data by the information acquisition module 1 are as follows:

[0079] Step 1, performing data cleaning on the historical browsing data, such as missing value processing, outlier processing, duplicate data processing, data consistency check, etc., to obtain a first data set;

[0080] Step 2, performing data conversion on the first data set, such as standardization, normalization and smoothing, to obtain a second data set;

[0081] Step 3, performing feature extraction on the second data set, such as principal component analysis and clustering algorithm, to obtain a third data set;

[0082] Step 4: perform data integration on the third data set, such as RTSP / RTMP / RTP streaming protocol integration and cloud video platform integration, to obtain a fourth data set.

[0083] It should be noted that by cleaning the historical browsing data, errors, inconsistencies and redundancies in the historical browsing data can be eliminated, and the quality of the historical browsing data can be improved. By converting the first data set, the impact of the dimension can be eliminated and the noise can be reduced, making the data easier to process and analyze. By extracting features from the second data set, it is helpful to identify key information in the data, reduce unnecessary complexity, and improve the efficiency and interpretability of the model. By integrating the data, it is convenient for subsequent analysis and application of the data, thereby enhancing the availability and flexibility of the data.

[0084] The specific steps of the data processing module 2 processing and calculating the current input data and the historical browsing data to obtain the adaptation value of each video resource in the historical browsing data and the current input data are as follows:

[0085] Step 1, obtaining the search command currently input by the learner;

[0086] Step 2, using a text recognition algorithm to extract features from the search instruction to obtain a demand feature vector;

[0087] Step 3, obtaining video feature vectors corresponding to all video resources in the fourth data set;

[0088] Step 4, calculating the similarity value between each video resource feature vector and the demand feature vector, and screening the similarity values ​​to obtain a first similarity value of the same type;

[0089] Step 5, summing the similarity values ​​of the first category to obtain a total similarity value;

[0090] Step 6: The adaptation value of the video resource is calculated by formula (1):

[0091]

[0092] in represents the adaptation value of the video resource, n represents the total number of similarity values ​​of the first type, and n is a positive integer. Indicates the first similarity value of the video feature vector.

[0093] By providing the data processing module 2, the present invention can screen the learner's historical browsing data according to the learner's current input features, and find out the historical browsing data similar to the learner's current input features, thereby improving the accuracy of subsequent video recommendations.

[0094] Calculate the similarity value between each video feature vector and the required feature vector, and filter the similarity values ​​to obtain the first similarity value of the same type, which is:

[0095] Step 41, calculate the similarity value between each video resource feature vector and the demand feature vector by formula (ii), the formula is as follows:

[0096]

[0097] where cosθ i represents the cosine value between each video feature vector and the required feature vector, θ i represents the angle between each video feature vector and the required feature vector, A represents the required feature vector, B i represents the video feature vector;

[0098] Step 42, compare the similarity value with a preset first similarity threshold interval. If the similarity value belongs to the first similarity threshold interval, mark the similarity value as a first similarity value of the same type; if the similarity value does not belong to the first similarity threshold interval, discard the similarity value and the video resource corresponding to the similarity value.

[0099] The video evaluation module 3 evaluates the candidate video resources in the video library to obtain the score value of the candidate video resources, specifically:

[0100] Step 1: preprocess the candidate video resources in the video library and obtain the feature vectors of the candidate video resources;

[0101] Step 2, calculating the similarity value between the feature vector of the candidate video resource and the required feature vector, and screening the similarity values ​​to obtain a second similarity value of the same type;

[0102] Step 3: Calculate the score of the candidate video resource using formula (III), which is as follows:

[0103]

[0104] where Df k Indicates the score value of the candidate video resource corresponding to the second similarity value. Represents the second similarity value of the selected candidate video resources, Cr k Represents the click rate of the selected candidate video resources, Sh k Pj represents the number of shares of the selected candidate video resources. k represents the feedback score value of the screened candidate video resource, ξ1 represents the similarity weight between the candidate video resource and the current input data, ξ2 represents the click rate weight of the candidate video resource, ξ3 represents the sharing weight of the candidate video resource, and ξ4 represents the feedback score weight of the candidate video, satisfying ξ1+ξ2+ξ3+ξ4=1, and ξ1 is greater than any one of ξ2, ξ3 and ξ4.

[0105] By providing a video evaluation module 3, the present invention can match candidate video resources similar to the learner's current input data in the video library, thereby further improving the accuracy of subsequent video recommendations. At the same time, by calculating the score values ​​of the screened candidate videos, it also prepares for subsequent video recommendations.

[0106] The similarity value between the feature vector of the candidate video resource and the required feature vector is calculated, and the similarity value is screened to obtain the second similarity value of the same type, which is specifically:

[0107] Step 21, calculate the similarity value between the feature vector of the candidate video resource and the required feature vector through formula (IV), the formula is as follows:

[0108]

[0109] where cosα k Represents the cosine value between the feature vector of the candidate video resource and the required feature vector, α k represents the angle between the feature vector of the candidate video resource and the required feature vector, A represents the required feature vector, H k A feature vector representing a candidate video resource;

[0110] Step 22, compare the similarity value of the candidate video resource with a preset second similarity threshold interval. If the similarity value belongs to the second similarity threshold interval, mark the similarity value as a second similarity value of the same type; if the similarity value does not belong to the second similarity threshold interval, discard the similarity value and the candidate video resource corresponding to the similarity value.

[0111] The video recommendation module 4 evaluates the adaptation value of each video resource in the historical browsing data and the current input data and the score value of the candidate video resource, and calculates the recommendation score of the candidate video resource in the video library, as follows:

[0112] Step 1: Obtain the feature vector B of the video resource corresponding to the first similarity value of the same typei ;

[0113] Step 2: sort the second similarity values ​​in order of magnitude, and extract the feature vectors H of the candidate video resources corresponding to the top n second similarity values ​​in order of absolute value from large to small. i ;

[0114] Step 3: Calculate the eigenvector H i With the eigenvector B i and marked it as indirect similarity;

[0115] Step 4: Get the feature vector H i The indirect adaptation value of the corresponding candidate video resource;

[0116] Step 5: The recommendation score of the candidate video resources in the video library is calculated by formula (V), which is as follows:

[0117]

[0118] where Df i Indicates the score value of the candidate video resources corresponding to the top n second similarity values. It represents the indirect adaptation value of the candidate video resources corresponding to the top n second similarity values, β1 represents the score weight, β2 represents the indirect adaptation weight, and β1+β2=1, β1 is greater than 1.5β2.

[0119] Calculate the eigenvector H i With the eigenvector B i The similarity is marked as indirect similarity and is implemented by formula (six), which is as follows:

[0120]

[0121] where cosγ i Indicates the indirect similarity between the feature vector of the candidate video resource corresponding to the second similarity value of the top n and the video feature vector, γ i Indicates the angle between the feature vector of the candidate video resource corresponding to the top n second similarity values ​​and the video feature vector.

[0122] Get the eigenvector H i The indirect adaptation value of the corresponding candidate video resource is realized by formula (VII), which is as follows:

[0123]

[0124] Where n represents the total number of first similarity values.

[0125] By providing a video recommendation module 4, the present invention can evaluate the adaptation value of each video resource in the historical browsing data and the current input data and the score value of the candidate video resources, and calculate the recommendation score of the candidate video resources in the video library, so that when the platform makes recommendations to learners, the learner's historical browsing data can also be considered, thereby improving the accuracy of recommending videos to learners.

[0126] The visualization display module 5 recommends and displays the candidate video resources with the highest recommendation scores in the video library to the learner as follows:

[0127] Step 1: sort the recommendation scores of the candidate video resources corresponding to the top n second similarity values ​​in the video library;

[0128] Step 2, obtaining the candidate video resource corresponding to the largest recommendation score, and recommending the candidate video resource to the learner;

[0129] Step 3: Track learners’ feedback scores, sharing times, and click-through rates of the candidate video resources in real time and update the feedback score weight, sharing weight, and click-through rate weight to achieve subsequent accurate recommendations for candidate video resources.

[0130] The present invention is provided with a visualization display module 5, so that after recommending a video to a learner, the platform can track the learner's feedback score, sharing times and click-through rate of the candidate video resource in real time and update the feedback score weight, sharing weight and click-through rate weight, so as to achieve subsequent accurate recommendation of the candidate video resources.

[0131] The above description of the embodiments enables those skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. An educational service resource cloud computing platform, characterized in that: include: The information acquisition module (1) is used to acquire the learner's current input data and the learner's historical browsing data, and pre-process the historical browsing data, and then send the current input data and the historical browsing data to the server for storage; A data processing module (2) is used to process and calculate the current input data and the historical browsing data to obtain an adaptation value between each video resource in the historical browsing data and the current input data; The video evaluation module (3) is used to evaluate the candidate video resources in the video library to obtain the score value of the candidate video resources, specifically: Preprocess the candidate video resources in the video library and obtain feature vectors of the candidate video resources; Calculate the similarity value between the feature vector of the candidate video resource and the required feature vector, and screen the similarity values ​​to obtain a second similarity value; The score of the candidate video resource is calculated by formula (III), which is as follows: where Df k Indicates the score value of the candidate video resource corresponding to the second similarity value. Represents the second similarity value of the selected candidate video resources, Cr k Represents the click rate of the selected candidate video resources, Sh k Pj represents the number of shares of the selected candidate video resources. k represents the feedback score value of the screened candidate video resources, ξ1 represents the similarity weight between the candidate video resources and the current input data, ξ2 represents the click rate weight of the candidate video resources, ξ3 represents the sharing weight of the candidate video resources, and ξ4 represents the feedback score weight of the candidate video, satisfying ξ1+ξ2+ξ3+ξ4=1, and ξ1 is greater than any one of ξ2, ξ3 and ξ4; The similarity value between the feature vector of the candidate video resource and the required feature vector is calculated, and the similarity value is screened to obtain the second similarity value of the same type, which is specifically: The similarity value between the feature vector of the candidate video resource and the required feature vector is calculated by formula (IV), and the formula is as follows: where cosα k Represents the cosine value between the feature vector of the candidate video resource and the required feature vector, α k represents the angle between the feature vector of the candidate video resource and the required feature vector, A represents the required feature vector, H k A feature vector representing a candidate video resource; Comparing the similarity value of the candidate video resource with a preset second similarity threshold interval, and if the similarity value belongs to the second similarity threshold interval, marking the similarity value as a second similarity value of the same type; If the similarity value does not belong to the second similarity threshold interval, the similarity value and the candidate video resource corresponding to the similarity value are discarded; A video recommendation module (4), used to evaluate the adaptation value of each video resource in the historical browsing data and the current input data and the score value of the candidate video resource, and calculate the recommendation score of the candidate video resource in the video library; The visualization display module (5) is used to recommend and display the candidate video resources with the highest recommendation scores in the video library to the learners.

2. The educational service resource cloud computing platform according to claim 1, characterized in that: The steps of preprocessing the historical browsing data by the information acquisition module (1) are as follows: Performing data cleaning on the historical browsing data to obtain a first data set; Performing data conversion on the first data set to obtain a second data set; Performing feature extraction on the second data set to obtain a third data set; Perform data integration on the third data set to obtain a fourth data set.

3. The educational service resource cloud computing platform according to claim 2, characterized in that: The data processing module (2) processes and calculates the current input data and the historical browsing data to obtain the adaptation value of each video resource in the historical browsing data and the current input data. The specific steps are as follows: Obtain the search command currently input by the learner; Using a text recognition algorithm to extract features from the search instruction to obtain a demand feature vector; Obtaining video feature vectors corresponding to all video resources in the fourth data set; Calculate the similarity value between each video resource feature vector and the demand feature vector, and filter the similarity values ​​to obtain a first similarity value of the same type; Sum the similarity values ​​of the first category to obtain a total similarity value; The adaptation value of the video resource is calculated by formula (1): in represents the adaptation value of the video resource, n represents the total number of similarity values ​​of the first type, and n is a positive integer. Indicates the first similarity value of the video feature vector.

4. The educational service resource cloud computing platform according to claim 3, characterized in that: Calculate the similarity value between each video feature vector and the required feature vector, and filter the similarity values ​​to obtain the first similarity value of the same type, which is: The similarity value between each video resource feature vector and the demand feature vector is calculated by formula (2), and the formula is as follows: where cosθ i represents the cosine value between each video feature vector and the required feature vector, θ i represents the angle between each video feature vector and the required feature vector, A represents the required feature vector, B i represents the video feature vector; Compare the similarity value with a preset first similarity threshold interval, and if the similarity value belongs to the first similarity threshold interval, mark the similarity value as a first similarity value of the same type; If the similarity value does not belong to the first similarity threshold interval, the similarity value and the video resource corresponding to the similarity value are discarded.

5. The educational service resource cloud computing platform according to claim 4, characterized in that: The video recommendation module (4) evaluates the adaptation value of each video resource in the historical browsing data and the current input data and the score value of the candidate video resource, and calculates the recommendation score of the candidate video resource in the video library, as follows: Obtain the feature vector B of the video resource corresponding to the first similarity value i ; Sort the second similarity values ​​in order of size, and extract the feature vectors H of the candidate video resources corresponding to the top n second similarity values ​​in order of absolute value from large to small. i ; Calculate the eigenvector H i With the eigenvector B i and marked it as indirect similarity; Get the eigenvector H i The indirect adaptation value of the corresponding candidate video resource; The recommendation score of the candidate video resources in the video library is calculated by formula (V), which is as follows: where Df i Indicates the score value of the candidate video resources corresponding to the top n second similarity values. It represents the indirect adaptation value of the candidate video resources corresponding to the top n second similarity values, β1 represents the score weight, β2 represents the indirect adaptation weight, and β1+β2=1, β1 is greater than 1.5β2.

6. The educational service resource cloud computing platform according to claim 5, characterized in that: Calculate the eigenvector H i With the eigenvector B i The similarity is marked as indirect similarity and is implemented by formula (six), which is as follows: where cosγ i Indicates the indirect similarity between the feature vector of the candidate video resource corresponding to the second similarity value of the top n and the video feature vector, γ i Indicates the angle between the feature vector of the candidate video resource corresponding to the top n second similarity values ​​and the video feature vector.

7. The educational service resource cloud computing platform according to claim 6, characterized in that: Get the eigenvector H i The indirect adaptation value of the corresponding candidate video resource is realized by formula (VII), which is as follows: Where n represents the total number of first similarity values.

8. The educational service resource cloud computing platform according to claim 7, characterized in that: The visualization display module (5) recommends and displays the candidate video resource with the highest recommendation score in the video library to the learner as follows: Sort the recommendation scores of the candidate video resources corresponding to the top n second-class similarity values ​​in the video library; Obtain the candidate video resource corresponding to the largest recommendation score, and recommend the candidate video resource to the learner; Track learners' feedback scores, sharing times and click-through rates of the candidate video resources in real time and update the feedback score weight, sharing weight and click-through rate weight to achieve subsequent accurate recommendations for candidate video resources.

Images