REGRESSION-BASED VARIANCE ESTIMATORS FOR THE ERROR COMPONENTS MODEL

In this paper, the authors propose a class of variance estimaotrs for the error components models which arises from applying the least squares principle. The suggested approach offers a new alternative when facing negative variance estimates and deals with unbalanced settings in a very natural way where no observation need be discarded. Finally, the least squares based variance estimaotrs can handle general decompositions of the error term (such as nested structures and heteroscedastic components, for example) in a very straightforward way. (A)

  • Corporate Authors:

    CENTRE DE RECHERCHE SUR LES TRANSPORTS. UNIVERSITE DE MONTREAL

    C.P. 6128, SUCCURSALE A
    MONTREAL, QUEBEC  Canada  H3C 3J7
  • Authors:
    • Bolduc, D
    • Laferriere, R
  • Publication Date: 1990

Language

  • English

Media Info

Subject/Index Terms

Filing Info

  • Accession Number: 00674395
  • Record Type: Publication
  • Source Agency: Transportation Association of Canada (TAC)
  • Files: ITRD
  • Created Date: Mar 8 1995 12:00AM