A Causality Mining Method Based on Dropout Behavior in MOOC Data

Abstract

The invention discloses a method for establishing a causal relationship model of dropout behavior based on MOOC data, by analyzing MOOC log data sets and constructing candidate independent variables that affect dropout behavior, and qualitatively analyzing the correlation between the candidate independent variable and the dependent variable property; design a quantitative measurement method for the dependence between the candidate independent variable and the dependent variable to construct an undirected graph, obtain an undirected graph composed of node sets composed of independent variables and dependent variables, and use local causality based on mutual information The network structure discovery algorithm eliminates the wrong variables based on the regression analysis equation in the undirected graph and generates a local network based on the conditional independence test to construct a directed local causal network structure for dropout behavior. For any target learning effect variable, The construction of the local causal network structure of the learning effect can be carried out through undirected graph generation, error node elimination and local network structure construction, and the causal relationship mining of the learning effect can be carried out.

Description

technical field [0001] The invention belongs to the technical field of artificial intelligence, and in particular relates to a causal relationship mining method based on dropout behavior in MOOC data. Background technique [0002] MOOC (Massive Open Online Course, Massive Open Online Course) provides a new type of teaching mode for the majority of learners through the online learning platform. It provides a completely free teaching method for a large number of users who need online learning. Educational resources and convenient learning experience, dedicated to large-scale course and content sharing, fragmented course form, and interaction among various groups, etc., MOOC has received more and more attention in the society. The rise of the MOOC platform has caused a great stir in the academic world, and domestic and foreign scholars such as Coursera, edX, and Udacity have also focused their attention on the massive data of the MOOC platform. The data was analyzed in detail ...

AI Technical Summary

Problems solved by technology

Although according to the basic assumption, the learners with fewer event operations on the MOOC platform are more likely to drop out of school, but how to use strict causal reasoning to construct a local causal network structure and determine the direction in the structure, It is still a very difficult problem

[0008] 3. Two variables that appear to be correlated may exhibit spurious causality

Images