Application of hybrid adaptive filters for stock market prediction

Prediction of stock market trends has been an area of great interest both to those who wish to profit by trading stocks in the stock market and for researchers attempting to uncover the information hidden in the stock market data. Traditional techniques such technical analysis and signal processing techniques such as moving averages and regression have had limited success in predicting markets, which could be attributed to the dynamic behavior of the markets. In signal processing, adaptive filters have been widely used for efficient filtering of signals. However, the utilization of adaptive filters for prediction, especially of financial signals, has not received much attention in literature. In this study, hybrid adaptive filters are introduced for prediction to obtain highly accurate results. The hybrid filters used are DCT-LMS, DCT-NLMS, DCT-RLS and Kalman filters. The proposed method is used to predict the values of five of the largest stock markets, namely, BSE100, NASDAQ, NIKKEI225, S&P NIFTY, and FTSE100. The performance of hybrid adaptive filters is compared against the conventional filters like autoregressive (AR), Moving Average (MA) filters and adaptive filters like LMS, NLMS etc. The base technique considered is the Random Walk (RW) process which acts as the benchmark technique. The results show a high degree of prediction accuracy for the hybrid adaptive filters, which is very high when compared to conventional filters, thus indicating that hybrid adaptive filters can be successfully used for stock market prediction.