ARIMA MODEL VERIFICATION WITH OUTLIER FACTORS USING CONTROL CHART

Abstract

Control charts are often used in quality control processes, especially in the industrial sector because of their significant benefits in increasing industrial production. However, control charts can also be used throughout the field of time series modeling to evaluate measures of accuracy represented by a particular time series model. The application of control charts in this research meets the criteria for evaluating accuracy. However, it is not certain that the time series model will have a high level of accuracy. There are various factors that can influence this phenomenon, one of which is the potential for outliers. Therefore, it is very important to perform time series modeling by adding an outlier factor. The residuals of the time series model obtained are used to create a control chart for model verification. The aim of this research is to evaluate the validity of time series models by looking at the influence of outlier characteristics to improve their accuracy. This research studies the accuracy of a time series model built using Gross Domestic Product (GDP) data in Indonesia. There are two different models, namely the ARIMA model without outlier factors and the ARIMA model with outlier factors which are used for research purposes. Both models were performed using the same data set. The results of this study indicate that the ARIMA model with outlier factors has better accuracy than the ARIMA model without outlier factors. This conclusion can be drawn based on the observation that the residual value is within the predetermined control limits, thus indicating that the process is in a state of statistical control.