Andrew Kroger - Backward Elimination Multivariate Stochastic Modeling (A case study of arsenic contamination of

groundwater), Advisor Dr. Malla

In statistics, stochastic modeling is a widely used technique in a variety of fields. Backward elimination is a particular technique of stochastic modeling that can be used in a multivariate setting. This project looks at the systematic development of the theory and methods behind this modeling, and includes a case study. Properties of multi-collinearity and homoscedasticity have been tested for the model developed. The interactive stochastic model to determine 'the amount of arsenic present in drinking water' by using its 15 covariates will be discussed.