Modification of Box-Jenkins methodology by injecting genetic algorithm technique

Abstract

The Box-Jenkins(BJ) methodology has four stages in modeling forecast time series data. The stages are model identification, model estimation, model validation and model forecast. The difficulties in modeling BJ is determining the right order in model identification and identifying the right parameter in model estimation. This study, genetic algorithm (GA) is proposed to solve the problem of model identification and model estimation. International tourist arrival to Malaysia is used as a case study to illustrate the effectiveness of this proposed model. The forecast result generated from this proposed model outperform single BJ model