METHOD AND APPARATUS FOR RECOMMENDING TEACHING AND LEARNING DATA USING MACHING LEARNING

摘要

The present disclosure provides a method for recommending teaching and learning materials using machine learning, when a plurality of users inquire (click) the teaching and learning materials, information (user_ID) that can identify the user's information and the corresponding teaching and learning materials can be identified extracting two or more pieces of information including information (data_ID) and processing it in the form of a data frame (a) (S1-1); Information (data_ID) that can identify the teaching/learning materials contained in the corresponding user information (user_ID) by extracting the corresponding information of each user information (user_ID) from the data frame (a) obtained in step S1-1 and the corresponding professor After performing one or more clustering operations using the Euclidean Distance method based on the analysis value after dimensionality reduction of the information (data_ID) that can identify the learning material, the representative value (c) of each cluster is extracted step (S2-1); Separately from steps S1-1 and S2-1, data is uploaded whenever a specific event occurs (e.g., periodic time, upload of new teaching and learning materials in the platform, etc.) apart from data frame (a) generation and representative value (c) extraction Information (data_ID) that includes more than 6 types of information including the title, registrant, subject, grade, usage, number of views, etc. of teaching and learning materials on the page within the platform After extracting the crawling method and processing it in the form of a JSON file, processing it in the form of a data feature data frame (b) (S1-2); Each piece of information corresponding to the data feature data frame (b) processed in step S1-2 is reduced to two or more feature values by using the method of principal component analysis (PCA). In this case, in the process of using the principal component analysis, a certain variation may occur in the principal component analysis method according to a preset criterion. After going through this process, generating a data frame (d) including information (user_ID) for identifying the data and the analysis result value (S2-2); Using the representative value (c) of each user's cluster extracted in step S2-1 and the data frame (d) generated in step S2-2, the representative value (c) for each user and the similarity of each teaching/learning material were calculated using the Euclidean distance. (Euclidean Distance) method is used to generate a similarity measurement value (e), and based on the similarity measurement value (e), a recommendation vector ( f) generating (S3); and recommending a predetermined number of materials from among a plurality of materials to at least one of the plurality of users based on the recommendation vector (f) (S4); it is about