A COMPARISON OF EFFICIENCY ON INTERVAL ESTIMATION OF VARIANCE ON NORMAL AND CONTAMINTED NORMAL DISTRIBUTIONS BY RESAMPLING TECHNIQUES
Keywords:
Normal Distribution, Contaminated Normal Distribution, Sample Variance, Jackknife, Standard BootstrapAbstract
The objective of this research is to compare the efficient confidence interval of variance based on normal and contaminated normal distribution by sample variance method and resampling techniques consisted of jackknife and standard bootstrap methods. The data sample are generated from the normal distribution with the parameter of mean () 2 and variance () 2 and 6. The mean and variance of contaminated normal distribution are the same as the normal distribution, but proportion of contamination () is 0.1 and scale factor () is 2 and 5. The sample sizes () for this study are set as 10, 20, 30 and 50, and the 95% and 99% confidence intervals.The simulated data is generated by R program and repeated 1,000 times in each situation. The criterion to consider the best confidence interval is a minimum value of average width. On normal distribution, the results are shown that sample variance method is the best confidence interval estimation. For contaminated normal distribution, jackknife and standard bootstrap methods outperform the sample mean method, and the standard bootstrap method has the average width narrower than jackknife method.
Downloads
References
Prachoom Suwattee. (2010). Theory of Statistical Inference. 3th ed. Bangkok: WVO officer of Printing Mill.
John R. Taylor. (2013). An Introduction to Error Analysis the Study of Uncertainties in Physical Measurements. (Jirapong Kasivitamnuay, Translator). Bangkok: Chulalongkorn University Printing House.
Quenouille, M.H. (1956, December). Notes on Bias in Estimation. Biometrika, 43, 353-360.
Efron, B. (1979, January). Bootstrap Methods: Another Loot at the Jackknife. The Annals of statistics, 7, 1-26.
Sornsawan Boonpen, Boonorm Chomtee, and Apinya Hirunwong. (2015, October-December). A Comparison of Intervals Estimation Methods for Scale Parameter of the Two-Parameter Weibull Distribution. Thai Science and Technology Journal Thammasat University, 23(4), 579-587.
Bumrungsak Phuenaree, and Kornrawee Rungsawang. (2017). An Estimation of Exponential Distribution Median by Bootstrap Methods. In 9th Science Research Conference. MA 187-194. Chonburi: Burapha University.
Anuwat Khamma, Manachai Rodchuen, and Putipong Bookkamana. (2017, October-December). Confidence Interval Estimation for the Ratio of Two Population Variances with Non-normal Distributions. KKU Research Journal (Graduate Studies), 17(4), 12-23.
Bonett, D.G. (2006). Confidence Interval for a Ratio of Variance in Bivariate Nonnormal Distribution, Journal of Statistical Computation and Simulation, 76(7), 637-644.
Sangnawakij, P., and Niwitpong, S. (2010). Interval Estimation for the Common Coefficient of Variation of Gamma Distributions. Thailand Statistician, 18(3), 340-353.
Efron, B., and Tibshirani, R.J. (1993). An Introduction to the Bootstrap. New York: Chapman & Hall.
Downloads
Published
How to Cite
Issue
Section
License
Srinakharinwirot University Journal of Sciences and Technology is licensed Under a Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International (CC-BY-NC-ND 4.0) License, Unless Otherwise Stated. Please Read Journal Policies Page for More Information on Open Access, Copyright and Permissions.
