文件列表:

Poster.pdf (353486, 2023-11-27)
ProjectReport.pdf (500057, 2023-11-27)
chromedriver-win64/ (0, 2023-11-27)
chromedriver-win64/chromedriver-win64/ (0, 2023-11-27)
chromedriver-win64/chromedriver-win64/LICENSE.chromedriver (244932, 2023-11-27)
chromedriver-win64/chromedriver-win64/chromedriver.exe (16863232, 2023-11-27)
final_stock_market_analysis_prediction_using_lstm.ipynb (1985249, 2023-11-27)
output_new.png (92802, 2023-11-27)

# Game-Theory-Based-Stock-Price-Prediction We started off a baseline LSTM model, and we are inserting different biases - 1. News sentiment bias, 2. Game Theory Bias, 3. User Hunch Bias. # Stock Price Predictor with LSTM and Game Theory Biases ## Overview This project combines advanced machine learning techniques, including Long Short-Term Memory (LSTM) networks and Game Theory-based biases, to predict stock prices. Additionally, sentimental analysis of news headlines and major global events is incorporated to provide a comprehensive understanding of market dynamics. ## Table of Contents - [Introduction](https://github.com/suyamoonpathak/Game-Theory-Based-Stock-Price-Prediction/blob/master/#introduction) - [Key Features](https://github.com/suyamoonpathak/Game-Theory-Based-Stock-Price-Prediction/blob/master/#key-features) - [Methodology](https://github.com/suyamoonpathak/Game-Theory-Based-Stock-Price-Prediction/blob/master/#methodology) - [Results](https://github.com/suyamoonpathak/Game-Theory-Based-Stock-Price-Prediction/blob/master/#results) - [Usage](https://github.com/suyamoonpathak/Game-Theory-Based-Stock-Price-Prediction/blob/master/#usage) - [Installation](https://github.com/suyamoonpathak/Game-Theory-Based-Stock-Price-Prediction/blob/master/#installation) - [Contributing](https://github.com/suyamoonpathak/Game-Theory-Based-Stock-Price-Prediction/blob/master/#contributing) - [License](https://github.com/suyamoonpathak/Game-Theory-Based-Stock-Price-Prediction/blob/master/#license) ## Introduction In financial markets, predicting stock prices is a challenging task. This project aims to enhance prediction accuracy by combining LSTM networks with Game Theory-based biases and sentimental analysis of news headlines. The inclusion of major event biases further refines the model, offering a holistic approach to stock market prediction. ## Key Features - **LSTM-Based Prediction:** Utilizes LSTM networks for capturing intrinsic statistical trends in time series stock data. - **Game Theory-Based Biases:** Incorporates biases derived from speculators' optimal strategies, assessing risk appetite and its influence on market behavior. - **Sentimental Analysis:** Analyzes news headlines for sentiment biases, offering insights into external factors impacting stock prices. - **Major Event Biases:** Considers major global events like the COVID-19 pandemic, geopolitical conflicts, and employee layoffs, assessing their impact on market sentiments. ## Methodology - **Data Collection:** Gathers stock prices from Yahoo Finance API and integrates major event data from reliable sources. - **LSTM Training:** Trains the LSTM model on historical stock data to learn intrinsic statistical trends. - **Game Theory Biases:** Calculates biases based on speculators' optimal strategies and risk assessments. - **Sentimental Analysis:** Scrapes news headlines and performs sentiment analysis to quantify external sentiment biases. - **Major Event Biases Integration:** Incorporates major event biases into the predictive model. ## Results ### Before Incorporating Biases: - RMSE: 5.45 - MSE: 18.95 - MAE: 3.81 - R2 Score: 0.89 ### After Incorporating Biases: - RMSE: 3.29 - MSE: 10.87 - MAE: 2.36 - R2 Score: 0.99 ## Usage 1. Clone the repository. 2. Install the required dependencies. 3. Follow the provided notebooks or scripts for training and prediction. ## Installation ```bash pip install -r requirements.txt

近期下载者：

相关文件：

评论：[我要评论] [举报此文件]

收藏者：