Forecasting notebook computer price as a function of constituent features

This paper proposes a three-phase approach to forecast a competitive price for a notebook computer as a function of constituent features. Phase I uses regression analysis to relate computer price to constituent features in each of a series of time periods. Phase II involves a time series analysis of each regression coefficient to quantify how trends in market conditions (e.g., the evolution of component technology) affect the market value of each feature. Phase III uses the time series analysis to forecast future market values of each feature and combines these results to forecast a competitive selling price for a notebook model that is composed of a selected set of features. The approach can be used in support of management decisions related to prescribing when to upgrade a notebook model and what features to include in each upgrade. The approach can be used, for example, to forecast notebook model price at introduction, and the rate at which price will erode over the model’s life cycle. Computational results indicate that the approach can forecast the price of a notebook computer model up to four months in advance of its introduction to within 10%. It can also forecast the rate of price erosion to within 10% for up to seven months after introduction — the length of the life cycle of a typical notebook model. Since this approach uses only publicly available data, it can be used easily in industry.