STOCK MARKET ANALYSIS AND PREDICTION USING MACHINE LEARNING MODELS

Abstract

This research addresses the challenge of accurately forecasting stock market trends, a critical task for investors and financial analysts. By employing a dual approach that integrates machine learning and traditional economic models, this study assesses stock price trends using both technical and sentiment-based indicators. Machine learning models, including deep learning methods such as LSTM and traditional models like Ridge, Decision Tree, Random Forest, and SVR, are compared with the economic model, Geometric Brownian Motion (GBM). The methodology involves thorough data preprocessing, model training, hyperparameter tuning, and evaluation across key metrics-including R-squared, MAE, MSE, RMSE, and execution time-to examine the relative performance of each model. Findings highlight that while the LSTM and CNN+LSTM model excel at capturing intricate patterns in stock data, traditional economic models offer interpretability and speed, which can be advantageous in certain contexts. Incorporating both technical and sentiment-based indicators is shown to enhance predictive accuracy across both model types, contributing a comprehensive framework for financial forecasting and comparison. This study offers valuable insights into the strengths and limitations of both machine learning and economic approaches within stock market analysis.