الفهرس 
الغلافOnly 14 pages are availabe for public view
 |  | from | 76 |  |  |
| | | from | 76 | | |
Abstract
Mortality and morbidity from respiratory disorders are among the
most frequent causes of death and morbidity worldwide. Study
after study has demonstrated that computer-assisted learning tools
improve self-directed learning, as well as improving problemsolving and forecasting skills Only a small number of respiratory
medicines have been produced so far, though. Thus, in this study
a framework has been developed for recognizing respiratory
illnesses using Data Mining Techniques (DMT) on clinical info.
In order to evaluate the prediction model, the framework consists
of three key phases: preprocessing, data mining, and evaluation.
Feature selection algorithm has been applied on the clinical data.
Mostly a 20 percent sample was used to test the model. The
remaining 80 percent was used to train a classifier. The resulting
classes were about the diagnosis of the respiratory system disease,
i.e., Normal Spirometry, Moderate Restriction, Mild Restriction
and Severe Restriction. The decision tree has 83.3% sensitivity,
92.1% specificity, the Classifier’s overall accuracy is 90.9%. As
part of the reasoning phase, a tool has been developed to execute
the prediction process.