B. Building the model and analysing for forecasting
The factors that affect the stability of palm oil are analysed
models for forecasting. The procedure is as follows.
1) Take the data to analyse the stationary by testing the
unit root based on Augment Dickey and Fuller
2) Take the stationary data at the same level both
dependent variables and independent variables (at level 1st
moment and/or 2nd moment) and analyse the long-term
relationship or find the co-integration. If variables in the
model have a long-term relationship at the same level, it can
be used to build the best model.
3) Estimate the model to build The Best Model, namely
independent variables have to show the true affect on
dependent variables. Considering the tau-statistics which have
to be statistically significant at a given level (level %5, %16
and )%15each.
4) Take The Best Model after that finish testing the 3
types of problems, namely Auto correlation (LM-test),
Heteroskadasticity (ARCH–Test) and Multicorrelinearity.
Then, respond to the correlogram value compared with the
chi-square which has to be checked with Ramsey Reset Test to
ensure the reliability of the model, and then can be utilized for forecasting the future and so on.