Title: Demand forecasting of tea by seasonal ARIMA model

Authors: E.V. Gijo

Addresses: SQC & OR Unit, Indian Statistical Institute, 8th Mile Mysore Road, Bangalore – 560059, India

Abstract: A tea packaging company in India was implementing supply chain planning process to improve its delivery performance. For this purpose the company was interested in forecasting the monthly demand for tea from its depots across the country. Time series data on demand of tea for 57 months were available. This series was modelled by Box-Jenkins seasonal auto regressive integrated moving average (ARIMA) model. Adequacy of the fitted model has been tested using Ljung-Box test criteria followed by residual analysis. Thus, the most appropriate model was used to forecast the monthly demand of tea. This model has helped the organisation to plan the production activities more efficiently so that shortages or excess production can be avoided.

Keywords: demand forecasting; Box-Jenkins model; seasonal ARIMA models; autoregressive integrated moving average; auto correlation function; ACF; autocorrelation; partial auto correlation function; PACF; Ljung-Box test; statistical testing; forecast accuracy; Akaike|s information criterion; AIC; augmented Dickey–Fuller test; ADF; tea packaging; India; SCM; supply chain management; delivery performance; monthly demand; supply depots; time series data; residual analysis; production planning; shortages; excess production; business excellence.

DOI: 10.1504/IJBEX.2011.037252

International Journal of Business Excellence, 2011 Vol.4 No.1, pp.111 - 124

Published online: 27 Sep 2014 *

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