Adaptive Response Rate Exponential Smoothing Forecast in AnalYzing IBM's Revenue

Abstract

Identification of appropriate forecasting method for time series data is a difficult job for most managers as forecasts are often inaccurate and uncertain. This paper explores several forecasting techniques in analyzing IBM's quarterly sales data for finding an appropriate method. Analysis of forecasts error has been performed for best method identification. Comparisons have been made on the basis of forecast accuracy measures and tracking signal test. It was found that adaptive response rate exponential smoothing (ARRES) technique is the optimal and best predictive technique.