12/18/06

Welcome to Thomas's Web site!

Thomas Shih is an Industrial Engineering Ph.D. at UTA. The title of his dissertation is "Convex Versions of Multivariate Adaptive Regression Splines and Implementations for Complex Optimizations Problems," co-supervised by Dr. Chen and Dr. Kim.

Below are the frequent questions about Shih’s research experience:

Q1: What project have you developed involving knowledge discovery through data mining, predictive modeling, and similar technique?

One good example will be the deicing project sponsored by Dallas Fort Worth International Airport.

Q2: What kind of analytical task did you perform?

In the DFW project, data mining approaches such as regression trees models were extensively applied to help the team effectively predict the response variable DO (dissolved oxygen) in the receiving waters of the Deicing project.

Q3: What was your major task in the Deicing project?

I was mainly responsible for aspects of processing deicing data from receipt of data to creation of analytic data sets and reports used in the Deicing project.

Q4: How did you handle the difficulty in the process of creating the Deicing relational databases?

I solved the problem independently through the help of various resources such as library and internet.

Q5: Have you ever trained or taught the other team members or clients how to create the databases?

I trained my teammates and enjoyed the knowledge sharing experience in applications and presentations in the Deicing project. It was a pretty rewarding process to me.

Q6: Can you give an example how you worked under pressure and in adverse situation?

Documenting the research from the dissertation gave me a lot of pressure. To over the adverse situation, I had to divide and conquer. Finally, it turned out that I could finish the job one by one.

Q7: Talk about your experience using relational databases and a database query language.

The Deicing project is based on a relational MS-ACCESS database. In order to perform the predictive analyses, queries had to be constructed to obtain various data sets on demand.

Q8: What data mining tools or predictive modeling tools did you use in the Deicing project.

The modeling tools I used include Regression, ANOVA, CART, MARS, FDR-based Multiple Comparisons Tests, and PCA.

This site was last updated 12/18/06