The Case Study of Forest Bioenergy as a Method for Selecting the Best Forecasting Model
Insights into Economics and Management Vol. 10,
28 May 2021
,
Page 96-104
https://doi.org/10.9734/bpi/ieam/v10/2202F
Abstract
The purpose of this article is to investigate technique for finding the best model for estimating wood fuel need in Greece for the years 2020, 2025, and 2030, with the aim of making forest bioenergy decisions. A complete time series of historical data exists that concerns a) the consumption of fuelwood and b) the six most important from the independent variables that could influence the consumption of fuelwood, which data cover the time period 1989-2010. The evaluation and choice of the best model was realized with the help of the following six statistical criteria: a) the size of standard error of theoretical values of dependant variable, S.E., b) the value of adjusted R square, , c) the non-existence of autocorrelation among the residuals (?i) through the criterion Durbin-Watson, d) the statistical significance of models coefficients through t criterion, e) the statistical significance of models through F criterion and f) the non-existence of multicolinearity through the values of Variance Inflation Factor. The price of fuelwood influences negatively the consumption of fuelwood, namely as the price of fuelwood increases the consumption of fuelwood decreases. The price of heating oil influences positively the consumption of fuelwood, namely as the price of heating oil increases the consumption of fuelwood increases.
- Evaluation of models
- choice of the best model
- statistical criteria