|
Activity Number:
|
146
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, August 12, 2002 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Risk Analysis*
|
| Abstract - #301780 |
|
Title:
|
Using Data Mining Techniques to Identify Volatile Organic Compounds Associated with Asthma Attack
|
|
Author(s):
|
Shiying Wu*+ and Jun Liu and Ye Hu and Edo Pellizzari
|
|
Affiliation(s):
|
RTI International and RTI International and RTI International and RTI International
|
|
Address:
|
3040 Cornwallis Road, Durham, North Carolina, 27709, u.s.a.
|
|
Keywords:
|
asthma ; chromatogram ; data mining ; exposure ; VOC
|
|
Abstract:
|
Volatile organic compounds (VOCs) are a group of important air pollutants. The purpose of this study is to use data mining technology to explore the differences in patterns of VOCs found in breath samples collected from asthma patients at baseline and onset of asthma attacks. Method: Longitudinal breath samples were collected from 24 asthmatic children. Samples were analyzed using gas chromatograph/mass spectrometry. Cluster analysis was performed to identify VOCs detected by chromatograms. Pattern recognition techniques were applied to the data to identify clustering among the VOCs and the correlation of their presence and intensity to the onset of the asthma attacks. Results: Interesting clusters of VOC compounds were discovered as potential triggers of asthma attacks. A group of compounds that might be used for potential chemical markers of the asthma attack were also identified. Significance: Chromatograms are routinely generated in analytical chemistry labs yet this rich information resource has long been overlooked in heath research community. Our results indicate that data mining has great potential in analytical chemistry as well as health researches.
|
