Campus card data-based bad learner prediction method

Abstract

The invention discloses a poor learning person prediction method based on campus card data. The method comprises the following steps: extracting the behavior features of students according to the records, recorded in the campus card data of college students, of class, library entering, restaurant consumption, bathroom consumption and commerce consumption; on the basis, unified mathematical representation is carried out on school activities of students in a semester, and then a hybrid network model based on the combination of a convolutional neural network and a multi-layer long-term and short-term memory network is provided to construct a poor learning person prediction model. According to the invention, the problem of early prediction of poor learners in colleges and universities is solved, the poor learners can be found in time in the semester, and targeted intervention and guidance are provided.

Description

technical field [0001] The invention relates to a method for predicting poor learners based on campus card data. Specifically, a hybrid network model combining a convolutional neural network and a multi-layer long-short-term memory network is used to predict poor learners. Background technique [0002] Prediction of poor learners is one of the important research contents in the field of educational data mining and learning analysis. In the traditional education environment, in recent years there have been studies on the prediction of poor learners based on student campus card data. [0003] The student's campus card data lasts from the beginning of the semester to the end of the semester, covering transaction data, library access control data, classroom attendance data, etc. during the semester. Among them, the transaction data records the consumption behavior of students in supermarkets, bathrooms, restaurants and other places; the library access control data records the t...

AI Technical Summary

Problems solved by technology

[0004] Existing related research usually extracts students' living characteristics and learning characteristics from the campus card data based on statistical methods, such as the number of times they eat breakfast, the number of times they go to and from the library, the number of times they borrow books, and the number of times they take baths every week, etc., and then predict the poor learners. Few studies have used deep learning methods to predict poor learners based on campus card data

Images