Title: A forecasting framework for the Indian healthcare sector index

Authors: Jaydip Sen

Addresses: Department of Data Science and Artificial Intelligence, Praxis Business School, Bakrahat Road, Off Diamond Harbor Road, Kolkata 700104, West Bengal, India

Abstract: Forecasting of future stock prices is a complex and challenging research problem due to the random variations that the time series of these variables exhibit. In this work, we study the behaviour exhibited by the healthcare sector's time series of India in the Bombay Stock Exchange (BSE). We collect the historical monthly index values of the BSE S&P healthcare sector from January 2010 to December 2021. The time series is decomposed into its three components trend, seasonality, and random. The component values reveal some important characteristics of the sector in the pre-pandemic and peri-pandemic times. We also propose five predictive models based on the exponential smoothing and autoregressive integrated moving average techniques for forecasting the monthly index values of 2021 based on the historical index values from January 2010 to December 2020. Extensive results are presented on the performances of the models.

Keywords: time series decomposition; trend; seasonality; randomness; exponential smoothing; HoltWinters forecasting; autoregressive integrated moving average; ARIMA; root mean square error; RMSE; mean absolute percentage error; MAPE.

DOI: 10.1504/IJBFMI.2022.125783

International Journal of Business Forecasting and Marketing Intelligence, 2022 Vol.7 No.4, pp.311 - 350

Received: 07 Mar 2022
Accepted: 12 Mar 2022
Published online: 28 Sep 2022 *

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