Analysis of Stock Market Behaviour by Applying Chaos Theory

Today, the results obtained from machine learning represent a substantial increase in returns over existing algorithmic trading engines. What we are proposing is an approach that is modeled around Chaos Theory. As we know, it is the study of surprises, of randomness. We aim to find occurrences of fractal data in huge chunks of stock market behaviour thereby effectively looking to have a better understanding of a certain stock's unpredictable behaviour. The existing approaches aim at maximising the gains and minimising the losses only, which does not enable us to comprehend the movement of the price index. This paper puts forth a novel idea that makes an effort in just that direction. We use normalisation-denormalisation and pattern mining to achieve our purpose.