A novel framework for stock trading analysis using casual relationship mining

Stock market has been around as a platform for buyers and sellers of shares. Success in stock trading, which depends on comprehensive knowhow on the trends of trading, is the dream of thousands of enthusiasts. Investors employ different means of analysis for expert decision making including periodical trade indicators, industry growth indicators, and economic forecasts. An automated approach for stock trading analysis is inevitable as there is huge amount of temporal data is involved. Several techniques came into existence as revealed in the literature. Data mining techniques are, of late, widely used for discovering trends of stock market. High-confidence rules extracted using such techniques are found instrumental to make well-informed decisions. In this paper we propose a framework with an underlying approach to have casual data mining which explores inter-transaction relationships that are latent in the stock datasets. This is used to unearth the trends in stock trading that can help investors to have strategies for profitable business. Our framework also provides placeholders for government policy and unexpected incidents to reduce error rate in the prediction of stocks. Our empirical results reveal that the proposed framework improves stock trading analysis and prediction accuracy significantly.