Journal of applied mathematics & informatics
- 제26권1_2호
- /
- Pages.177-189
- /
- 2008
- /
- 2734-1194(pISSN)
- /
- 2234-8417(eISSN)
GOODNESS OF FIT TESTS BASED ON DIVERGENCE MEASURES
- Pasha, Eynollah (Department of Mathematics, Teacher Training University) ;
- Kokabi, Mohsen (Department of Mathematics, Central Tehran Branch, Islamic Azad University) ;
- Mohtashami, Gholam Reza (Department of Mathematics, Birjand University)
- 발행 : 2008.01.30
초록
In this paper, we have considered an investigation on goodness of fit tests based on divergence measures. In the case of categorical data, under certain regularity conditions, we obtained asymptotic distribution of these tests. Also, we have proposed a modified test that improves the rate of convergence. In continuous case, we used our modified entropy estimator [10], for Kullback-Leibler information estimation. A comparative study based on simulation results is discussed also.