Authors: Jaydip Sen; Tamal Datta Chaudhuri
Addresses: Department of Information Technology and Analytics, Praxis Business School, Bakrahat Road, Off Diamond Harbor Road, Kolkata 700104, West Bengal, India ' Department of Finance and Economics, Calcutta Business School, Diamond Harbor Road, Bishnupur 743503, 24 Parganas South, West Bengal, India
Abstract: The objective of this work is to ascertain sectoral characteristics of stock market indices of India through time series decomposition. It is postulated that different sectors in an economy do not behave uniformly and sectors differ from each other in terms of their trend patterns, their seasonal characteristics and also their randomness. The R programming language is used for decomposing the time series of the sectoral indices as well as indices of stocks. Based on our approach, we also provide a framework of forecasting that can be applied on various sectors of an economy and also on individual stocks in the sectors. Results indicate that the behavioral patterns of stocks largely follow those exhibited by the respective sectors to which they belong. The forecasting methods are found to be robust and accurate in forecasting the pattern of index movements of the sectors and the individual stocks.
Keywords: time series decomposition; trend; seasonality; random; forecasting; holt winters forecasting method; ARIMA; auto correlation; auto regression; moving average; partial auto correlation; portfolio management.
International Journal of Business Forecasting and Marketing Intelligence, 2018 Vol.4 No.2, pp.178 - 222
Received: 29 Jun 2017
Accepted: 19 Sep 2017
Published online: 14 Feb 2018 *