지원사업
학술연구/단체지원/교육 등 연구자 활동을 지속하도록 DBpia가 지원하고 있어요.
커뮤니티
연구자들이 자신의 연구와 전문성을 널리 알리고, 새로운 협력의 기회를 만들 수 있는 네트워킹 공간이에요.
논문 기본 정보
- 자료유형
- 학위논문
- 저자정보
- 지도교수
- 백창룡
- 발행연도
- 2015
- 저작권
- 성균관대학교 논문은 저작권에 의해 보호받습니다.
이용수1
이 논문의 연구 히스토리 (2)
- 이 논문의 후속연구가 궁금하신가요?
- 연관 학술논문 또는 학술발표를 통해 보다 발전된 연구결과를 확인하실 수 있습니다.
- 이 논문의 연구 히스토리 확인하기
초록· 키워드
오류제보하기
It is well known that the vector autoregressive (VAR) models in high dimension
suffer from noisy estimates, unstable predictions and hard interpretation. To overcome
such shortcomings, the so-called sparse vector autoregressive (sVAR) model
which forces many small coefficients in VAR to exactly zero has been suggested and
shown to be very effective. This paper studies coupling measures to select non-zero
coefficients in sVAR. The basic idea is based on the simulation study reveals that
removing the effect of other variables greatly improves the performance of coupling
measures. Also, the coefficients of sVAR model is asymmetric, hence asymmetric
coupling measures such as Granger causality certainly improve computational
costs. Therefore, this paper propose two asymmetric coupling measures based on
the filtered residuals series, namely filtered-cross-correlation and filtered-Grangercausality.
It is shown that the proposed coupling measures work quite well even for
heavy-tailed and high order sVAR models in the simulation study and real data.
suffer from noisy estimates, unstable predictions and hard interpretation. To overcome
such shortcomings, the so-called sparse vector autoregressive (sVAR) model
which forces many small coefficients in VAR to exactly zero has been suggested and
shown to be very effective. This paper studies coupling measures to select non-zero
coefficients in sVAR. The basic idea is based on the simulation study reveals that
removing the effect of other variables greatly improves the performance of coupling
measures. Also, the coefficients of sVAR model is asymmetric, hence asymmetric
coupling measures such as Granger causality certainly improve computational
costs. Therefore, this paper propose two asymmetric coupling measures based on
the filtered residuals series, namely filtered-cross-correlation and filtered-Grangercausality.
It is shown that the proposed coupling measures work quite well even for
heavy-tailed and high order sVAR models in the simulation study and real data.
목차
제1장 서론제2장 희박벡터자기상관회귀모형제 1절 벡터자기상관회귀 모형제 2절 변수 선택을 위한 상관 통계량2.1 부분스펙트럼 일관성2.2 그래인저 인과관계제3장 기존 방법론의 한계 및 필터링의 필요성제4장 필터를 이용한 커플링 측도제 1절 필터링된 잔차에 기반을 둔 교차 상관관계제 2절 필터링된 잔차에 기반을 둔 그래인저 인과관계제5장 모의실험제 1절 모의실험 모형1제 2절 모의실험 모형2제6장 실증 데이터제7장 결론
최근 본 자료
댓글(0)
0