The volatility of stock markets makes prediction a complex but essential task for investors seeking to optimise their decision-making. This study presents a data-driven decision support model for stock market prediction using historical stock price data of Infosys Ltd., a leading IT firm in India. By applying data mining techniques—specifically classification and rule-based prediction—the model aims to identify meaningful patterns from past trends and assist in forecasting future price movements.
The study uses monthly trading data from the National Stock
Exchange for the period 2010 to 2015, including open, high, low, and close
price values. These values were transformed into symbolic categories
(rise/fall) for analytical clarity. Using the ESTARD Data Miner tool, decision
trees and rule-based classifiers were generated to derive actionable prediction
rules. The rules were further tested using a What-If analysis for real-time
prediction scenarios.
The findings demonstrate the effectiveness of symbolic conversion
and decision tree modelling in predicting the stock trend classes. This
predictive framework holds potential for guiding investors in making more
informed buy/sell decisions and enhancing the reliability of investment
strategies based on historical data patterns.
Author
(s) Details
Sanjeev Gour
Medicaps University, Indore, MP, India.
Sanjana Sharma
Acropolis Institute of Technology & Research, Indore, MP, India.
Prerita Kulkarni
Medicaps University, Indore, MP, India.
Shimna Mohan K.
Medicaps University, Indore, MP, India.
Please see the book here:- https://doi.org/10.9734/bpi/nabme/v7/5193
No comments:
Post a Comment