1 AIT Asian Institute of Technology

Parameter estimation for the three-parameter lognormal distribution from complete and censored samples

AuthorHu, Peng
Call NumberAIT Thesis no. CS-87-8
Subject(s)Lognormal distribution
Hydrologic models
NoteA thesis submitted in partial fulfillment of the requirements for the degree of Master of Engineering, School of Engineering and Technology
PublisherAsian Institute of Technology
AbstractIn t his work, several methods of parameter estimation for the three-parameter lognormal (LN3) distribution have been studied under the condition of small (sample size ~ 100), complete samples. A new method, Sundry Average Method (SAM) , was developed. The comparisons of t he estimated parameters among different methods were carefully taken. Statistical descriptors for each method were also searched. As for censored samples, a type I censoring was considered and three cases, the right, l eft and double censoring, were investigated. Parameter estimation was carried out by using the method of maximum likelihood (MML) . Fisher's information matrix was derived for evaluating asymptotically the variances and covariances of the parameters estimators. Computer simulations were designed and corresponding computer programs was developed. A large number of replications (around 9500 effective replications for a definite sample size), for the sample size from 30 to 100 with an increment of 10, were generated to investigate the sampling properties of the estimators censored. It was found that the SAM provided estimators of t he T-year event with small bias, but their variances were comparable with the fv111, estimators only for small samples (size ~ 50) . Moreover, censoring to t he left seems to reduce the relative bias when low levels of censoring. The same seems to be true a l so for right censoring at high censoring points.
Year1987
TypeThesis
SchoolSchool of Engineering and Technology (SET)
DepartmentOther Field of Studies (No Department)
Academic Program/FoSComputer Science (CS)
Chairperson(s)Huynh Ngoc Phien
Examination Committee(s) Aagesen, Finn Arve ; Khang, Bach Hung
Scholarship Donor(s)DAAD, The Federal Republics of Germany
DegreeThesis (M.Eng.) - Asian Institute of Technology, 1987


Usage Metrics
View Detail0
Read PDF0
Download PDF0