In this presentation, I will introduce the "log-linear variance model" that can be used to fit models in JMP. Normally, when fitting a JMP model to perform regression analysis or analysis of variance (general linear model), equal variances are assumed. However, the data we actually handle often deviates from this assumption. The log-linear variance model is a very useful method in such cases. Several methods, such as variable transformation, are used to deal with unequal variances. In contrast, the JMP log-linear variance model analyzes the data as is, analyzes (models) how factors affect the mean and variance, and allows you to understand the results simply by using a prediction profiler. I will also introduce examples of how it is used at our company.
He graduated from the Graduate School of Materials Engineering at Nagoya Institute of Technology in March 1994 and was employed by a chemical manufacturer. He specialized in organic synthesis before changing jobs. After joining Gore Japan LLC, he committed himself to statistics and has been involved in statistical education and consultation for over 20 years. He encountered JMP shortly after he began his commitment to this work, and became convinced that there was no better statistical software for engineers to use. Since then, he has promoted the use of JMP as a statistical tool for use within the company. He has a strong interest in practical statistics, and continues to think about ways to utilize statistics using JMP in a wide range of fields, from research and development to production technology.
