Stock market prediction using various machine learning models

This repo contain code and implementation for Stacked LSTM, Logistic Regression, Random Forest, Naïve Bayes, Linear Support Vector Machine and Non-Linear Support Vector Machine.

Method

We have used NLP using historical News data with these algorithms: Logistic Regression, Random Forest, Naïve Bayes, Linear Support Vector Machine and Non-Linear Support Vector Machine.

And we have done Time-seris analysis with current data using Stacked LSTM model.

Requirements

Python Anaconda(preffered), tensorflow v2, pandas, numpy, keras, sklearn, datetime, matplotlib, seaborn Account in Tiingo for API key (only used in Stacked LSTM model)

Data Scraping

Historical News Data for NLP

We use The Guardian's API to gather the historical News data to be used for NLP for all algorithms (except Stacked LSTM).

Recent Stock data Data

In order to gather Recent/Historical stock data (like open, close, date, high, low, volume), we use Tingoo's API to scrape the data.

Prediction

Run the desired .ipynb according to the algorithm you need to implement.

Result

Here is the Accuracy graph of our project:

License

The underlying code of this project is licensed under the MIT license.