Abstract: Comparison of a forecast model, one of the essential elements of forecast model fitting, is totally a different concept than many a times confused concept of its accuracy measure. With accuracy measure alone, no any context or reference could be given to convene an investigated model is better or superior than any other model made from the same data set. This paper aims to demonstrate how a multiple regression model, constructed from time series data can be compared through naïve forecast approach. This, at the same time, has clarified the differentiation between comparison of a forecast model and its accuracy measure. Regarding the context of comparing the models, in this case, in both accuracy measures, mean absolute error and mean absolute percentage error, multiple regression model has taken the lead to be smaller error, meaning that fitted model of the choice i.e., the multiple regression model, was better to the naïve forecast, the benchmark method. This concludes it was worth fitting the multiple regression model. Keywords: forecast models, naïve method, forecast error, forecast accuracy, mean absolute error, mean percentage error. Title: A Naïve Approach for Comparing a Forecast Model Author: Chuda Prasad Dhakal International Journal of Thesis Projects and Dissertations (IJTPD) Research Publish Journals
