Description
The leasing business is one of the most important distribution channels for the automotive industry. This implies that forecasting accurate residual values for the vehicles is a major factor for determining monthly leasing rates: Either a systematic overestimation or underestimation of future residual values can incur large potential losses in resale value or, respectively, competitive disadvantages. In this paper, an operative DSS with the purpose of facilitating residual value related management decisions is introduced, with a focus on its forecasting capabilities. Practical implications are discussed, a multi-variate linear model and an artificial neural network approach are benchmarked and further, the effects of price trends and seasonal influences are investigated. The analysis is based on more than 150,000 data sets from a major German car manufacturer. We show that artificial neural network ensembles with only a few input variables are capable of achieving a significant improvement in forecasting accuracy.
Decision Support for the Automotive Industry: Forecasting Residual Values using Artificial Neural Networks
The leasing business is one of the most important distribution channels for the automotive industry. This implies that forecasting accurate residual values for the vehicles is a major factor for determining monthly leasing rates: Either a systematic overestimation or underestimation of future residual values can incur large potential losses in resale value or, respectively, competitive disadvantages. In this paper, an operative DSS with the purpose of facilitating residual value related management decisions is introduced, with a focus on its forecasting capabilities. Practical implications are discussed, a multi-variate linear model and an artificial neural network approach are benchmarked and further, the effects of price trends and seasonal influences are investigated. The analysis is based on more than 150,000 data sets from a major German car manufacturer. We show that artificial neural network ensembles with only a few input variables are capable of achieving a significant improvement in forecasting accuracy.