Title
Current Failure Prediction for Final Examination via Nearest Neighbor Method using Past Trends of Weekly Online Testing

Authors

Hideo Hirose


Source

International Journal of Learning Technologies and Learning Environments (IJLTLE), Vol 4, No 1, pp.**-**, ****, 2022


Abstract
We showed previously that we can predict the success/failure status for the final examination to each student at early stages in courses using the current trends of estimated abilities to the learning check testing in terms of the item response theory, where we used the same testing results in prediction and in construction of the mathematical model. However, such a treatment may cause the overfitting effect. In this paper, we have shown that we can still predict the current success/failure status for the final examination using the past trends of estimated abilities to the learning check testing and the past final examination results.
In prediction, we applied the nearest neighbor method for determining the similarity in the trends of estimated abilities to the learning check testing.

Key Words
current failure prediction, past trends, item response theory, nearest neighbor, similarity, online testing, learning analytics.

Citation

@

Times Cited in Web of Science:

Cited in Books: