Author(s)

A. Panagopoulos, Z. Psillakis & D. Kanellopoulos

Abstract

A new inferential model that allows low cost reliable time series forecasts is established. The model provides a new unique computationally straightforward approach based on widely used additive models. An application of the model referring to monthly percentages of occupancy in Greece is presented. The results are compared to those of the well know Box-Jenkins method. This study reveals the performance of the model and confirms its reliability. Keywords: time series, forecasting models. 1 Introduction The construction of time series models is an important aspect of business and economic analyses, because time series [1, 2] are many of the variables of most interest to business and economic researchers. Forecasting is about predicting the

Keywords