Method and device for recommending election courses

Abstract

The invention discloses a method and a device for recommending election courses. The method comprises the following steps: firstly, constructing the similarity between student course pairs based on students and course portraits, wherein a cosine similarity calculation mode is used for discrete attributes, a Gaussian kernel function calculation mode is used for continuous attributes, and a user course pair similarity matrix is obtained; then, based on improvement of an SVD algorithm and an RMSE loss function, decomposing the student course matrix into a student-characteristic moment-course characteristic matrix with course characteristics, and performing recombining to obtain a new student course matrix through analysis of preference degrees of students to the course characteristics and membership degrees of courses to the course characteristics, wherein the course matrix of the student contains the preference data of the student for the course; and finally generating the recommended course selection matrix according to the course matrix of the student, so that personalized recommendation of the course is made for the student, and the blindness of course selection of the student can be effectively solved.

Description

technical field [0001] The invention relates to the fields of big data analysis and artificial intelligence, in particular to personalized recommendation. Background technique [0002] After the reform of the college education system, colleges and universities have developed in the direction of multi-disciplinary comprehensive. The coverage of disciplines and majors continues to expand, and the number of courses offered is also increasing. The increase in the number of courses brings more choices for students to choose courses, but too many courses also make the students' course selection inevitably blind. [0003] At present, the extensive application of the teaching management system has accumulated a large amount of teaching practice data. However, there are many incompleteness in these data. For example, 25% of the graduates leave the school every year, and 25% of the new campus students will cause the proportion of cold start users. , loss ratio and other issues. The...

AI Technical Summary

Problems solved by technology

The increase in the number of courses brings more choices for students to choose courses, but too many courses also make the students' course selection inevitably blind

[0003] At present, the extensive application of the teaching management system has accumulated a large amount of teaching practice data. However, there are many incompleteness in these data. For example, 25% of the graduates leave the school every year, and 25% of the new campus students will cause the proportion of cold start users. , loss ratio and other issues

Difficulties and inaccuracies in the operation of the recommendation system caused by the defects of the data itself

For example, a teacher can be uniquely identified by ID in the teacher table, but only the teacher's name appears in other tables (such as the teacher's schedule table), and the ID is not used, which makes it impossible for the recommendation system to actually determine whether the course is Which teacher teaches

Images