A Comparison of Several Approaches to Missing Attribute Values in Data Mining
RSCTC '00 Revised Papers from the Second International Conference on Rough Sets and Current Trends in Computing
| Hi-index | 0.00 |
The paper describes the development of a clinical decision model to help Emergency Department physicians assess the severity of pediatric asthma exacerbations. The model should support an early identification (at 2 hours) of those patients who are having a mild attack and those who are having a moderate/severe attack. A comprehensive approach combining rough sets and expert-driven manual feature selection was applied to develop a rule-based decision model from retrospective data that described asthmatic patients visiting the Emergency Department. The experiment involved creating the following four potential decision models differentiated by the subsets of clinical attributes that were considered: Model A using all attributes collected in the retrospective chart study; Model B using only attributes describing the patient’s history; Model C using only attributes describing the triage and repeated assessments; and Model D using attributes from Model C expanded with some of the attributes from Model B identified by expert clinical knowledge. Model D offered the highest assessment accuracy when tested on an independent retrospective data set and was selected as the decision model for asthma exacerbations.