Title: Business health characterisation of listed Indian companies using data mining techniques

Authors: Senthil Arasu Balasubramanian; G.S. Radhakrishna; P. Sridevi; Thamaraiselvan Natarajan

Addresses: Department of Management Studies, National Institute of Technology, Tiruchirappalli-620015, Tamil Nadu, India ' Department of Management Studies, National Institute of Technology, Tiruchirappalli-620015, Tamil Nadu, India ' Department of Management Studies, National Institute of Technology, Tiruchirappalli-620015, Tamil Nadu, India ' Department of Management Studies, National Institute of Technology, Tiruchirappalli-620015, Tamil Nadu, India

Abstract: The purpose of our study was to predict the business health of listed Indian companies using data mining tools and algorithms called ANN-MLP and DT-QUEST and identify financial ratios that significantly affect the company's performance. We used 2,000 listed Indian companies with 12,000 firm-year records (cases) from 2011 to 2016 to predict the financial performances of the companies and classify them as successful or unsuccessful, based on 17 financial ratios. The final sample of data was divided into training and test set (50:50, 60:40, 70:30 and 80:20). The test results confirmed accuracy between 84% and 86% for the MLP technique and between 92% and 93% for the QUEST technique. Sensitivity analysis results showed that return on long-term fund, net profit margin, and operating margin are three critical variables that affect business health.

Keywords: business health; multilayer perception; sensitivity analysis; decision tree; financial ratios.

DOI: 10.1504/IJBIS.2019.10024676

International Journal of Business Information Systems, 2019 Vol.32 No.3, pp.324 - 363

Received: 13 Jun 2017
Accepted: 21 Jan 2018

Published online: 15 Oct 2019 *

Full-text access for editors Access for subscribers Purchase this article Comment on this article