Font Size: a A A

THE RATE OF WEAK CONVERGENCE OF A VECTOR OF U-STATISTICS GENERATED BY A SINGLE SAMPLE

Posted on:1982-08-12Degree:Ph.DType:Dissertation
University:University of California, BerkeleyCandidate:CARMICHAEL, WILLIAM PAULFull Text:PDF
GTID:1470390017465087Subject:Statistics
Abstract/Summary:
Let U(,n) denote a random vector in (//R)('k), k(, )(GREATERTHEQ)(, )1, consisting of kone-sample U-statistics where each of the k component U-statistics is generated by the same sample of size n. Under suitableregularity conditions on the associated kernel vector, it is shownthat the uniform rate of weak convergence of SQRT.(nU(,n) to the(, )(' )appropriate limiting normal probability measure over the class ofall Borel measurable convex subsets of (//R)('k) is of order 0(n('- 1/2)).;*This paper was prepared with the support of the U.S. Office ofNaval Research Contract No. N00014-75-C-0444.;...
Keywords/Search Tags:Vector, U-statistics
Related items