V1I3P2

Artificial Intelligence and Machine Learning in Predictive Stock Market Analysis

Amalendu Bhunia^1*

Abstract

This study explores the application of artificial intelligence and machine learning in predictive stock market analysis, utilizing robust econometric and statistical models. Secondary data spanning from 2015 to 2024 was sourced from Yahoo Finance, UCI Machine Learning Repository, RBI, NSE databases, and Sentdex, focusing on healthcare companies listed in the NSE database. The study employed a multiple regression model to examine the relationships between stock prices (dependent variable) and return on assets, sentiment analysis, unemployment rate, and the relative strength index. Results indicated significant positive effects of ROA and sentiment scores on stock prices, while unemployment rates had a negative impact. RSI emerged as a critical technical indicator, influencing momentum-driven price movements. For time-series forecasting, the ARIMA model was implemented to capture temporal dependencies in stock price movements. Findings revealed that past stock prices and forecast errors significantly influence future prices, with periodic patterns reflected in higher-order autoregressive terms.

Keywords:

Artificial intelligence; machine learning; stock market analysis; multiple regression analysis; ARIMA model.