Stocks prediction using machine learning and technical indicators
Abstract
Developing a predictive system for the stock market is important for algorithmic trading and investment management. Many researchers have used technical indicators to forecast stock markets and different financial markets. In this project, input window length and technical indicators are used as features to predict the stock market movement. The conducted experiments combined individual technical indicators to get the most important composite indicator for PSX (Pakistan Stock Exchange). Four technical indicators and ten stocks were used across eight different input windows during the experiments. Individual indicators gave better performance than the composite method when compared on the average, highest and lowest accuracies. Bollinger Bands turned out to be the best, followed by RSI (Relative Strength Index).
