Model Selection Criteria for Factor Analysis : Simulation Results
- 주제(키워드) 도움말 Model selection , factor model , Akaike's (1973) Akaike information criteria (AIC) , Hurvich and Tsay's (1989) corrected AIC (CAIC) , Baysian information criterion (BIC) , Hannan and Quinn's (1979) information criterion (HQc) , Wei (1992)'s Fisher information criterion (FIC)
- 발행기관 서강대학교 일반대학원
- 지도교수 최 인
- 발행년도 2013
- 학위수여년월 2013. 8
- 학위명 석사
- 학과 및 전공 도움말 일반대학원 경제학과
- 실제URI http://www.dcollection.net/handler/sogang/000000052607
- 본문언어 영어
- 저작권 서강대학교 논문은 저작권 보호를 받습니다.
초록/요약 도움말
This paper proposes information criteria for the selection of the number of static factors. These have conventionally been employed in regression and time series analyses, but have not been considered in factor models. As the first purpose of this paper, Akaike's (1973) Akaike information criterion (AIC), Hurvich and Tsay's (1989) corrected AIC, the Bayesian information criterion (BIC) of Akaike (1978) and Schwarz (1978), Hannan and Quinn's (1979) information criterion (HQc), and Wei's (1992) Fisher information criterion (FICκ) are formally derived for static factor models. The second purpose of this paper is to report simple simulation results that compare performance of extant and new criteria. The data generating process for the simulation consists of serially correlated factors, and i.i.d. and serially and cross-sectionally correlated idiosyncratic errors. The simulation results show that the criteria introduced in this paper are promising.
more