Earticle

초록

영어

This study evaluates the predictive performance of an extended Dynamic Factor Model (DFM) that incorporates port cargo throughput to reduce short-term information gaps caused by the publication lag of the Index of All Industry Production. Since the Index of All Industry Production is reported at a monthly frequency, it does not require the mixed-frequency aggregation or tent-loading restrictions typically used in quarterly GDP nowcasting. However, the ragged-edge problem remains because the release timing differs across input variables. To address this issue, this study applies a state-space DFM estimated using the Kalman filter and the Expectation-Maximization (EM) algorithm. The empirical results show that the Port-augmented DFM slightly reduces both RMSE and MAE compared with the benchmark DFM. The improvement is relatively more pronounced during the potential logistics shock period in 2021. However, since the overall improvement is limited and the Diebold-Mariano test is not conducted, the findings should be interpreted as exploratory evidence suggesting the complementary informational value of port cargo throughput.

목차

Abstract
Ⅰ. 서론
Ⅱ. 이론적 배경
2.1 전산업생산지수
2.2 DFM 기반 Nowcasting 방법론
2.3 산업생산지수 Nowcasting과 고빈도 데이터 연구
2.4 항만물동량 데이터의 활용
Ⅲ. 연구 설계
3.1 연구 분석 구조
3.2 변수 구성
3.3 정상성 검정과 변수 변환
3.4 동적 요인 모형
3.5 표본 분할과 축차 예측
3.6 EM 알고리즘
3.7 평가 지표
Ⅳ. 실증연구 결과
4.1 DFM 팩터로딩 분석
4.2 전산업생산지수의 추정 추이
4.3 예측오차 비교
4.4 결과 해석
Ⅴ. 결론
5.1 연구 결과
5.2 연구의 의의
5.3 한계 및 향후 연구
References

키워드

저자

• 이수민 [ 동아대학교 경영정보학과 ]
• 조용복 [ 동아대학교 경영정보학과 ]

참고문헌

발행기관

간행물

표지보기 관심저널 등록

• 간행물명

  한국경영정보학회 정기 학술대회 [KMIS Conference]

• 간기

  반년간

• 수록기간

  1990~2026

• 십진분류

  KDC 325 DDC 658