/CS3244-Predicting-Direction-of-Stock-Prices

Traditionally, stock price movements are hard to predict. Our project aims to predict the direction of stock prices based on financial data and news headlines. As part of our research, we compare and evaluate different classification models while concurrently exploring different natural language processing techniques to utilize news headlines for model improvement.

Primary LanguageJupyter Notebook

CS3244 Project - Predicting Direction of Stock Prices

  • The "Experiments" folder contains all our notebooks and NLP experiments.

  • The "NewYorkTimesAPI" folder contains the code for retrieving news headlines of companies (using the NewYorkTimes API).

  • The "S&P500 Dataset" contains the individual company stock data as well as the combined stock data over 5 years.