المقدرات المثلى لمعلمات التوزيع الطبيعي اللوغاريتمي باعتماد المحاكاة

Abstract

The statistical distributions possessed and still possesses the growing importance of the applied statistics and the diversity of distribution that can be owned is the log-normal which has important relation with one of the most important statistical distributions, a normal distribution characterized by the logarithmic normal distribution which represent many experiments especially in the fields of engineering, physics and agricultural. The aim of this study is to find the best method for estimating the distribution parameters, where different sample sizes and values of the parameters are generated by simulation with repetition in order to reach to the best estimation method stability.In the present study four ways of estimation are chosen, namely {Maximum likelihood estimation (MLE), Moment method (MOM), Jackknife Maximum likelihood estimation (JaMLE) and Jackknife Moment method (JaMOM)} and applied to the study data has shown results the ability of estimation. methods different access capabilities close to the parameters of the distribution is assumed as the results show possession of the estimation.The method (JaMLE) odds the biggest and by (48) percent for the first parameter(μ) while the results for the second parameter (σ) in the same way is the biggest by (63) percent, and this came high stability this method is characterized in providing the capabilities of being the closest to the true distribution parameters values.We may use other methods to estimate parameters distribution, such as (bayes, shrinkage), another size of samples and additional parameters values for the purpose of comparison to extract the best tactic.