Stock Price Time Series Forecasting: A Comprehensive Analysis of Traditional Machine Learning vs. Deep Learning Approach

Abstract

It is very important to correctly predict BSE Sensex stock prices because of the high market volatility and the increasing monetary value of investment decisions. Conventional forecasting methods have a tendency not to account for complex, non-linear patterns in the market. This paper attempts to forecast stock price movement through machine learning (ML) and deep learning (DL), using BSE Sensex 10- Year Stock Price data available at Kaggle. The data used is the daily data from April 2014 to March 2024 with open, high, low, and close prices of the index. The preprocessing of the data was used to deal with missing data, to transform the data types, and to chronologically order the records. StandardScaler was used to apply feature scaling in order to increase the performance of the model. Two predictive models, Random Forest (RF) and Long Short-Term Memory (LSTM), were implemented for stock price time-series forecasting. The models were compared on the basis of regression measures such as R², RMSE and MAE. The experimental findings indicate that the LSTM model outperforms the Random Forest model, with an R² of 0.9777, suggesting it is highly predictive. The results provide insight into the usefulness of deep learning (DL) methods for understanding temporal patterns and improving the accuracy of stock market forecasts.