Date of Award

Winter 1990

Project Type

Dissertation

Program or Major

Chemistry

Degree Name

Doctor of Philosophy

First Advisor

Sterling Tomellini

Abstract

The ever-increasing power of modern infrared instrumentation, coupled with the decreasing number of experienced spectroscopists has created an imbalance between information generation and interpretation capabilities. At the same time, digital computers are being developed which continue to grow in storage and processing capabilities, and shrink in cost. Clearly, the computer may serve as a valuable tool to aid the analytical chemist in interpreting spectroscopic information. This dissertation deals with the development of new approaches to exploiting computer technology to interpret infrared spectroscopic data.

A large existing expert system for functional group analysis, PAIRS, has been modified to transfer the maximum amount of information to the chemist. Two closely coupled knowledge based systems, IRBASE and MIXIR, have been created to identify major components of condensed phase mixtures. A second version of MIXIR has been developed to identify major components of vapor phase mixtures. Finally, a neural network approach to peak detection in analytical data has been developed.

Download

Share

COinS