Stock Market Prediction using Numerical and Textual Analysis

Objective

To create a hybrid model for stock price performance prediction using numerical analysis of historical stock prices, and sentimental analysis of news headlines.

Approach

• Extract Sentiment Scores from given news headlines dataset, with the help of nltk's SentimentIntensityAnalyzer.

• Use Multivariate Time Series Forecasting using the LSTM (Long Short-Term Memory) model in Keras and Tensorflow. LSTM analyses the features from both sentiment scores and numerical Historical stock data to predicit opening stock prices.

Results

• Achieved Training loss: 0.05 and Validation loss: 0.02

• Achieved RMSE on test data : 511

Data

• Historical stock prices(SENSEX (S&P BSE SENSEX)) from https://finance.yahoo.com/

• Textual (News) data from https://bit.ly/36fFPI6

References:

Deep learning for stock prediction using numerical and textual information- Ryo Akita, Akira Yoshihara, Takashi Matsubara, Kuniaki Uehara