Date of Award
8-25-2011
Document Type
Dissertation
Degree Name
Doctor of Philosophy (PhD)
Department
Information Science
First Advisor
Xiaowei Xu
Abstract
The ability to determine accurately the health of a fetus through non-invasive techniques has tremendous positive affects for the fetus, the mother, and the health care system. In this project, Knowledge Discovery in Databases (KDD) techniques were applied to fetal magnetocardiogram data from 118 human fetuses with a gestational age of 24 to 39 weeks. These techniques produced three models relating to the health of a fetus which are presented in this document. The first model distinguished maternal high-risk factors versus low-risk factors. The second model was used to classify neonate outcomes of sick versus healthy. The third model was used to classify fetuses with maternal high-risk factors yet born healthy versus sick neonates. These models were created with the support vector machine classifier using a combination of data from the time and frequency domains. The first model had a sensitivity of 0.67 and a specificity of 0.65. The second model's sensitivity and specificity values were 0.40 and 0.86 respectively. The third model had a sensitivity of 0.5 along with a specificity of 0.79.
Recommended Citation
Snider, Dallas H., "Knowledge Discovery in Fetal Activity Data" (2011). Theses and Dissertations. 285.
https://research.ualr.edu/etd/285
