Diagnose abnormal nasal based on the C4.5 modeling using cross section area curve from acoustic rhinometry

Abstract

Thisresearch proposes methods to classify the pattern of unusual nasal cavity using Ripper Rule, C4.5 decision tree, K-Nearest neighbor which aims to help physicians classify abnormal nasal cavity from acoustic rhinometry signal. The experiments showed that the algorithm was best effective classification is C4.5 decision tree has ROC 0.99 (sensitivity 0.99, specificity 0.99 and standard deviation 0.1). The result showed that abnormalities of the nasal cavity are about 0.3-5 cm. and nasal cross sectional area is less than 0.55 cm. 2. Therefore, this study suggests that the C4.5 decision tree algorithm could apply for screening abnormal nasal cavity. It led to application or tool development on medical devices in the future. © 2013 IEEE.