A Bayesian-adaboost model for stock trading rule discovery

摘要

Detecting the trading patterns with different technical indicators from the historical financial data is an efficient way to forecast the trading decisions in the financial market. In most cases, the trading patterns which consist of some specific combinations of technical indicators are significant in predicting the efficient trading decisions. However, discovering those combinations is a rather challenge assignment. In this paper, we propose a novel method to detect the trading patterns and later the Naive bayes with Adaboost method was employed to determine the trading decisions. The proposed method has been implemented on two historical stock datasets, the experimental results demonstrate that the proposed algorithm outperforms the other three algorithms and could provide a worthwhile reference for the financial investments.