Stock Prediction

For the Interactive Qualifying Project (IQP) at WPI, I collaborated with a team of 5 to create predictive models in Matlab that can make short-term predictions for stock prices. Historical stock prices were used with time-series analysis to make future predictions. Various methods and models were used including linear trend lines, Fourier analyses, error bounds, moving averages, normalization, and market influence. Other buy/sell signals were also investigated by analyzing volumes, Relative Strength Index (RSI), and Bollinger bands.

Each member of the team was responsible to test these models on our own chosen sectors. I was responsible for the Industrial sector and overseeing the coding component. 

Our predictive models were put to the test using a virtual stock trading platform, MarketWatch. Due to the volatility of the market, our models weren't too successful other than for a few days so sometimes, day trading is necessary to make a profit. Most of our sectors were still able to outperform the market.