__Mathematical Modeling Capabilities:__

Able to determine an appropriate way to represent a particular kindof reality using the vast array of possible mathematical and/or statisticaltechniques that are available.

Able to decide when and how to use deterministic models (such as thosedeveloped from differential equations), stochastic models (such as multivariateleast squares, logistic regression and related binary modeling techniques,principle components analysis, etc.), or models that combine deterministic and stochastic methods or are customized to reflect a specific situation.

__Experience:__

I developed a model for a client that formed the basis of a Company's response to a discrimination complaint. It was unlike anything in the textbooks, but was obviously more accurate in fitting the Company data than the model developed by the complainant. It clearly demonstrated that the complaint was based on a spurious relationship and was therefore without merit.

Many models are intended to forecast factors which will affect long-and short-range corporate planning and policy-making. Actual examples include:

- An attrition model which provided an estimate of employee terminations, retirements, etc. It was based on age, tenure, employee characteristics, and external economic factors.
- A salary estimation model which provided an estimate of each employee's salary was developed by inventing a special type of multiple regression. The model was based on salary level, tenure, education, occupation, organizational unit, etc. Each section of the model was developed so that it was mathematically independent of the other parts. In this way, what would have been a severe "multicollinearity" problem was neutralized. The result was a model inwhich the estimated salary was very highly correlated (r = .97) with actual salary across a group of over 5000 management employees.