PERBANDINGAN ORDINAL FOREST DAN REGRESI LOGISTIK ORDINAL

There are many methods used to identify the important predictors and predict the value of the ordinal response variable. However, currently, there are prediction methods for ordinal response variables that do not use the ordinal nature of the response variables. This study uses the ordinal forest method and as a comparison the ordinal logistic regression method. The accuracy and kappa value of the ordinal forest method in each replication (5 replications) is always greater than the ordinal logistic regression. Furthermore, the value of accuracy and kappa of each group based on GRDP in the ordinal forest method is always greater than the ordinal logistic regression. So that the ordinal forest method is better used in data on village development index in North Maluku Province in 2020.