Multivariate adaptive regression splines and insider trading data for stock prediction

Abstract

Data mining statistical techniques have not been used to improve the prediction of abnormal stock returns using insider trading data (except for one preliminary study by the author using neural networks). To expand data mining statistical research in this area, an investigation using Multivariate Adaptive Regression Splines (MARS) was initiated. The investigated covered 343 companies for a period of 4.5 years. Study findings revealed that the prediction of abnormal returns could be enhanced in the following ways : (1) extending the time of the future forecast up to 1 year ; (2) increasing the period of back aggregated data up to 4 months ; and (3) narrowing the assessment to certain industries such as electronic equipment and business services.