Hello, M. David says that the calculation of the mean you mentioned is valid when you know the real values of a lognormal distribution "a" and "b" parameters. In case you have a "lognormal" sample distribition, Sichel's t-estimator is the proposed alternative to estimate the population mean (M. David - Geostatistical Ore Reserve Estimation, pgs. 11- 19). Best regards, Jorge Carvalho
********************************************************************* Jorge M. C. M. Carvalho Departamento de Minas Faculdade de Engenharia da Universidade do Porto - FEUP R. Dr. Roberto Frias, s/n, 4200-465 Porto, Portugal Tel.: +351 225081992, Fax.: +351 225081448 E-mail: <mailto:[EMAIL PROTECTED]> [EMAIL PROTECTED] _____ From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Digby Millikan Sent: sábado, 30 de Dezembro de 2006 17:05 To: AI Geostats mailing list Subject: AI-GEOSTATS: Mean Is the Mean calculation M. David is talking about (m= e^a+b^2/2) for when the lognormal distribution in hand is a sampling population, and one is trying to estimate the mean of the entire population?
