Method and Apparatus for Stock Performance Prediction Using Momentum Strategy along with Social Feedback

Abstract

Stock prediction and historical stock data analysis have been of great interest over the decades. The research is wide from classical deterministic algorithms to machine learning models and techniques along with the supply huge amounts of historical data. Volatility and Market Sentiment are key parameters to account for during the construction of any stock prediction model. Commonly used techniques like the n-moving days average is not responsive to swings in the stocks and the information sent and posted online has made a huge effect on investors' opinions on the market, making these the two optimal parameters of prediction. Hence, we present an automatic pipeline that has 2 modules - N-Observation period momentum strategy to identify potential stocks and then a stock holding module that identifies market sentiment using NLP techniques.