MMO - AN IMPROVED ESTIMATOR FOR LOG-PEARSON TYPE-3 DISTRIBUTION

A mixed method combining the method of moments and the method of optim ization (MMO) was developed for estimating the parameters of the log-P earson type 3 (LP3) distribution. The MMO estimates the parameters of mean and standard deviation by the method of indirect moments (MIM) an d estimates the coefficient of skewness by minimizing both the relativ e root average square error (RRASE) and the relative average bias (RAB ). Both the predictive capability and descriptive capability of six po pular estimation methods were evaluated using 90 sets of observed floo d data and six selected LP3 populations with 1000 samples for each sel ected sample size. The performance of the MMO was compared with those of five other selected estimation methods. A weighted ranking index (W RI) procedure was developed to help select the best combination of dis tribution and method for the Louisiana flood data- The WRI takes both the predictive capability and the descriptive capability into account in the evaluation. The combination of LP3/MMO was found to be the best combination for Louisiana flood data.