Financial Sentiment Analysis for Stock Price Prediction

Overview

This repository contains code and datasets for exploring the correlation between sentiment analysis of financial text (tweets) and stock price prediction. The goal is to demonstrate how emotional sentiment from financial texts can impact stock market movements over time.

Purpose

• Establish a significant link between sentiment scores and stock price fluctuations.
• Evaluate sentiment's effect over different time frames.
• Use machine learning algorithms for price prediction.
• Validate using metrics like MSE, RMSE, R², and correlation coefficients.

Key Methods

• Sentiment Analysis:
  • RoBERTa: Transformer model for extracting sentiment scores.
  • VADER: Lexicon-based sentiment analysis tool.
• Machine Learning:
  • Regression and classification models to predict stock prices.
  • Various algorithms tested to assess accuracy and predictive power.

Data

• Tweets related to financial markets.
• Stock price data sourced from Kaggle for training machine learning models.

Validation Metrics

• Mean Squared Error (MSE)
• Root Mean Squared Error (RMSE)
• R² Score
• Pearson Correlation Coefficient

Applications

• Computational Finance: Understanding the impact of market sentiment on stock prices.
• Quantitative Analysis: Building predictive models based on financial data and sentiment analysis.

Keywords:

Sentiment Analysis, Stock Price Prediction, RoBERTa, VADER, Machine Learning, Time Series Forecasting, Financial Text, Kaggle Dataset, Regression Models, Classification Models, MSE, RMSE, R² Score, Pearson Correlation, Predictive Analytics, Natural Language Processing (NLP), Data Science, Stock Market Prediction, Financial Analysis, Algorithm Evaluation, Data Validation.