Method for automatically recommending topic contents

Abstract

The invention relates to the technical field of machine learning, and discloses a method for automatically recommending topic contents. The method that utilizes information accurate to an internal structure of a topic is utilized to perform calculation and can automatically complete topic recommendation precise to the knowledge point level is provided, that is, on the one hand, in the case of given data, the topic recommendation can be completed automatically without needing human intervention at all. On the other hand, by introducing detailed information of steps in the topic and the capability information of user details, the recommended topic content is more in line with the user's needs. In addition, since random factors are introduced in the method, the problem of repeated recommendation in the same scenario can be solved, and the selection can be randomly performed according to the degree of adaptation, which not only ensures the correlation between the selection of the topic andthe degree of adaptation, but also avoids the dilemma of push-down determination under the same scenario.

Description

technical field [0001] The invention belongs to the technical field of machine learning, and in particular relates to a method for automatically recommending topic content applicable to the education industry. Background technique [0002] In the education industry, recommending the most suitable practice questions to students can effectively improve the learning effect of students when the number of questions to be done is determined. [0003] In the traditional way, this part is generally completed by the instructor. The explaining teacher chooses the appropriate topic for the students to practice based on their own understanding of the students and the topic in their own impression. First of all, considering the number of questions that students have to do, it is difficult for teachers to directly recommend a new question after completing a question in real time. Secondly, considering the limited memory of teachers, it is difficult to remember each student's ability in ...

AI Technical Summary

Problems solved by technology

This method can ensure that the recommended topics meet the requirements of students from the general direction of the topic, but in fact the recommendation accuracy is still very poor

First of all, the classification-based scheme is not completely reasonable. The questions often involve knowledge from multiple aspects. Forcibly classifying them into a certain category will lose the accuracy of recommendation.

At the same time, since the difficulty is often marked for the topic, the matching of students' abilities can only be done at the topic level, and the interaction of difficulty between the various parts of the topic is often ignored.

Due to the above two reasons, the data accuracy of its recommendation algorithm can only reach the topic level, and its recommendation accuracy is often worse

[0006] At the same time, most of the traditional recommendation algorithms determine the questions based entirely on the size of the indicators, which makes it easier for students to see the same questions at the same level, and repeating the same question cannot effectively improve the students' real level, so it is completely The determined algorithm cannot meet the needs of students in solving problems

Images