Speaker: Stephen McCoy, UTK

Title: iQSPR methods for material design

Abstract: Establishing a relationship between the raw chemical products used in industrial polymers and the properties of their end product material is a time consuming and expensive endevour. Mathematical models address problems of this type and have aided research and development efforts across the whole spectrum of material design. Most of the modeling efforts to date have focused on solving quantitative structure-property relationships (QSPR) by starting from the known chemicals and establishing a mathematical model to predict the resulting properties.

In this talk I will cover the modeling procedure of QSPR inversion (iQSPR), where the starting point is the desired properties and a statistical learning model is employed to predict the chemical products and compositions that would give rise to those properties. The development, execution, and characteristic results of the model applied to industry data will be discussed as well as future implications from the prediction ability of the model.