Mathematical Sciences - Colloquium

Analysis of the Distribution of Grades from the Fall 2001 Calculus 152 Final Examination & Nonparametric Estimation of a Linear Relationship for Bivariate Data Professor Rudy Gideon The University of Montana The Correlation Principle for the estimation of a parameter theta: A statistical inference or procedure should be consistent with the assumption that any explanation of a set of data should be accompanied by theta-hat, a value of theta, that makes some correlation coefficient zero. Although extensive methods have been developed for linear models, generalized linear models, non-linear models, and time series models, as well as estimation of parameters for a particular distribution, we only have time to give one result. A nonparametric correlation coefficient measures monotonicity rather than linearity. It will be shown how to measure linearity with any nonparametric correlation coefficient. The Greatest Deviation correlation coefficient (GD) will be used and its robustness demonstrated. Thursday, 28 February 2002 4:10 p.m. in Math 109 Coffee/treats at 3:30 p.m. Math 104 (Lounge)

Spring 2002 Colloquium Schedule | Mathematical Sciences | The University of Montana