시계열 자료의 이상값 검출과 평활 방법

Identification of outliers and adjustment in time series data

It is well known fact that outliers should be identified and we adjust the effects of outliers for time series data to be analyzed properly. In this paper, a simple effective method to detect and adjust outliers is developed by using the distance measure between two data points. Moreover we develope a computer program of the proposed method developed by S-plus. And some ways of application for the suggested method are also discussed by analyzing a real empirical time series data.