Journal of the American Statistical Association, September 1st, 1994
The development of an estimation technique called adaptive mixtures is examined. The technique assumes that an estimate sequence converging to a true density generates asymptotically optimal performance in classification and discrimination problems. The adaptive mixture approach is developed from kernel estimation and finite mixture methods. The application of Monte Carlo simulations reveals the scope of the technique's application.
The estimation of a probability density function based on a sample [Mathematical Expression Omitted] of independent identically distributed
observations is essen...
HighBeam Research, Free Preview: 'Adaptive mixtures.'... Full Membership required for unlimited access. Free 7-day trial.
Subscribers: HighBeam content is only available to HighBeam subscribers. Click the link above for more information.
