Learning effect optimization method based on user behavior causal relationship in MOOC log data

The invention discloses a learning effect optimization method based on a user behavior causal relationship in MOOC log data. The method comprises the steps of based on log data of MOOC platform users,by means of user tendency calculation and matching, selecting and generating causal independent variables in a causal network according to a causal reasoning framework of a reverse fact, and obtaining the minimum data scale under different data scales according to the average Markov blanket length of the causal network when the length tends to be stable; obtaining a causal network group; screening edges among the network nodes by using an expert accuracy algorithm; grouping the screened network groups; comprehensively generating a final causal network by using a Bagging voting mechanism; obtaining the causal relationship between the user behavior and the learning effect; based on the causal relationship, reasonably planning the learning path of the user according to the reason variable node and the result variabl