TY - JOUR
T1 - Monte Carlo computation of the mean of a function with convex support
AU - Ritov, Y.
PY - 1989/2
Y1 - 1989/2
N2 - Let T=(T1,...,Tk) be a random vector, C a convex set and h(·) a function with C as its domain. In this paper the computation of E{h(T)|Tε{lunate}C}by Monte Carlo methods is considered. A Markov chain is constructed such that C is its sample space and its asymptotic distribution is the distribution of T given Tε{lunate}C. E{h(T)|Tε{lunate}C}is computed then as the time average of h applied on the chaim states.
AB - Let T=(T1,...,Tk) be a random vector, C a convex set and h(·) a function with C as its domain. In this paper the computation of E{h(T)|Tε{lunate}C}by Monte Carlo methods is considered. A Markov chain is constructed such that C is its sample space and its asymptotic distribution is the distribution of T given Tε{lunate}C. E{h(T)|Tε{lunate}C}is computed then as the time average of h applied on the chaim states.
KW - Isotonic regression
KW - Maximum likelihood
KW - Transformation models
UR - http://www.scopus.com/inward/record.url?scp=38249025178&partnerID=8YFLogxK
U2 - 10.1016/0167-9473(89)90027-3
DO - 10.1016/0167-9473(89)90027-3
M3 - ???researchoutput.researchoutputtypes.contributiontojournal.article???
AN - SCOPUS:38249025178
SN - 0167-9473
VL - 7
SP - 269
EP - 277
JO - Computational Statistics and Data Analysis
JF - Computational Statistics and Data Analysis
IS - 3
ER -