Stock Market Prediction model using TPWS and Association Rules Mining

The objective of this research is to classify or forecast the stock market from the general investor's point of view. There are three parts in this research. In the first part we performed a survey on most of the well known data mining indicators, implemented the algorithms and calculated the accuracy by applying them on historical data. Then we presented an indicator algorithm which has higher accuracy compare to existing algorithms and it also provides a decision point that helps the investor to understand the significance of the result of the indicator. Finally we applied association rules mining to group the selected (based on precision) indicator algorithms to come up with a model to increase the overall accuracy. However motivating fact is we achieved far better results from our suggested model than other comparable indicator algorithms or strategy. For our research we used the data of Dhaka Stock Exchange (DSE), capital market of Bangladesh.