Advances and Applications in Statistics
Volume 42, Issue 1, Pages 33 - 46
(September 2014)
|
|
A NEW NON PARAMETRIC METHOD OF ESTIMATION OF PREDICTION LIMITS AND COMPARISON OF ITS PERFORMANCE WITH SEMI PARAMETRIC METHOD
Ch. Lakshmi Sujatha and M. Vishnu Vardhana Rao
|
Abstract: The primary objective of the present work is to propose a non parametric method of estimation of prediction limits based on bootstrap samples and kernel smooth quantile estimator by Yang [14]. Comparison of the performance of the proposed method with a semi parametric estimation method GAMLSS is also made using data of Indian preschool children anthropometry extracted from NFHS-3. GAMLSS is a powerful statistical modelling technique which was used for deriving 2006 WHO growth standards for children. It was observed that there is no significant difference between the prevalence proportions of under weight, stunting and wasting for boys and girls between using GAMLSS and the proposed method BPLKSM local growth cut-offs for Indian children aged 0-36 months. Therefore, we can conclude that the method BPLKSM performs similar in assessment of nutritional status in comparison with GAMLSS. |
Keywords and phrases: non parametric estimation, kernel, prediction limits, bootstrap samples, semi parametric estimation, GAMLSS. |
|
Number of Downloads: 362 | Number of Views: 1350 |
|