Deep knowledge tracking method and system applied to multivariate programming exercises

The invention relates to a deep knowledge tracking method and system applied to multivariate programming exercises, and the method comprises the steps: receiving user exercise data and compiling log data of a programming platform system, and carrying out the data preprocessing; performing log mode clustering on the error information; constructing a question type feature matrix; position coding based on the time interval and the learning ability; performing one-hot coding conversion on the answer interaction sequence to obtain an answer interaction embedded vector; and inputting the answer interaction embedded vector into the trained deep knowledge tracking model to realize deep knowledge tracking. According to the method, through data processing, knowledge base establishment and model improvement, the model can be suitable for application of multivariate question types and better conforms to actual scenes, and particularly through establishment of the compiling result knowledge base, application of knowledge