HopHacks 2014 Project -- Sam Beckley, Rajiv Deshpande, Vignesh Ramchandran, Thomas Yi
Foreknown--A stock market predictor Android App that utilizes retrospective analysis to create a stochastic model for predicting future market behavior
Overview: In the modern stock market, financial enthusiasts from all fields are attempting to forecast the trends of a plethora of stocks. As the market has been around for decades now, it follows that many mathematical attempts to model the behavior of stocks exist; however, there has been a lack of exploration into the application of such models in the form of mobile applications. We consequently decided to implement our own flavor of a mathematical model that accomplishes the complex forecasting of fluctuating stocks within the convenience of an Android application.
Approach: We use a stochastic model to treat stocks as semi-random agents, bounded by drift and influenced by volatility, in order to make a prediction that uses a random, noise generating function to make future predictions as to how stocks will behave. Consequently, we graph the vectorized data from the mathematical model to display the predictions; as a sample, we have included the stock data for HP.
Future: We plan to launch this app on the marketplace free of charge for use by anyone, maintain the github repositiory, and potentially accept future pull requests from third parties. Additionally, we will work towards refining the application--in particular, we hope to implement a query system that can access stock prices from a market database.
Thank you for your time and interest!