Statistical Evaluation of Equivalence Test Based on the Genetic Diversity Index
碩士 === 國立成功大學 === 統計學系碩博士班 === 98 === The United Nations marked 2010 as the International Year of Biodiversity - the variety of life on Earth 2010. Biologists often define biodiversity as the "totality of genes, species, and ecosystems of a region". It is clear that the importance of bi...
| Main Authors: | , |
|---|---|
| Other Authors: | |
| Format: | Others |
| Language: | en_US |
| Published: |
2010
|
| Online Access: | http://ndltd.ncl.edu.tw/handle/98314974235622502544 |
| Summary: | 碩士 === 國立成功大學 === 統計學系碩博士班 === 98 === The United Nations marked 2010 as the International Year of Biodiversity - the variety of life on Earth 2010. Biologists often define biodiversity as the "totality of genes, species, and ecosystems of a region". It is clear that the importance of biodiversity. This study is concerned with the equivalence for multi-categorical data and focuses on gene comparison by some genetic diversity index, especially in nucleotide diversity index. The traditional estimator of diversity ignores the missing gene types result in to underestimate the gene diversity, when there is a non-negligible number of unseen genetic sequence. Therefore we combine the concept of Horvitz-Thompson and sample coverage to obtain the better genetic diversity estimator. Then we use the estimator and two one-sided test method to test the equivalence. A simulation study was conducted to empirically investigate the size and power of the proposed methods. Besides, a bootstrap-based approach is also proposed in small sample size and compared with the two one-sided test by the size and power.
|
|---|