Comparison of Forecasting Models

摘要

In this paper we compare two forecasting models applied to the large civil aircraft market share of the United States. The data used consists of Type 6 aircraft, which are commercial aircraft with a seating capacity between 171 to 240 passengers. The regression model applied to this data is an adapted logit multiregression type which has an r-squared of percent 75.2. The other model is a Box-Jenkins Auto regressive Integrated Moving Average, (ARIMA). This time series forecasting model considers the previous market share values to predict the future values. In this paper, a real world scenario is examined to compare these two models. The comparison basis will be the Mean Absolute Deviation (MAD) between actual and predicted values of the two models.