Characteristics of Multivariate Time Series and Review for Clustering Methods

• 발행기관: 한국상품학회
• 저자명: 최환석(Choi, Hwanseok) 김영일(Kim, Young‐ il)
• 간행물 정보: 『상품학연구』제29권 제3호, 95~105쪽, 전체 11쪽
• 주제분류: 경제경영 > 경제학
• 파일형태: PDF
• 발행일자: 2011.05.30

영문 초록

As technology for data storage in business has been improved, research for analyzing this complicated dataset has accounted for this matter. Data mining is one of these efforts. This study described the needs for statistical tool for clustering complicated multidimensional time series and representative statistical methods for clustering analysis including raw data or model based approach, feature extraction approach, and multi‐way principal component analysis (MPCA). As the closing this review study, Kernel variant principal component analysis (KPCA) was described as the alternativemethod for non‐linear cases.

목차

I. Introduction

II. Time Series Data

III. Time Series Clustering

IV. Review of Time Series Clustering Methods

V. Principal Component Analysis & Multi‐way Principal Component Analysis

VI. Discussion

References

키워드

multivariate time series MPCA clustering analysis kernelmethod