Time and Frequency Domain Performance Comparison for Wheeze Detection using K-Nearest Neighbor

摘要

In this paper, the comparison between the performance of wheezes data processing in the frequency domain and in the time domain is evaluated using K-Nearest Neighbor (KNN). The purpose of this paper is to clarify the confusion regarding the methods used nowadays, as many of the previous researchers have stated that wheezes data are better processed in the frequency domain due to its dominant frequency peaks but not a single researcher has made a direct comparison to prove the reliability of the method used. From the evaluation made, the result shows that the performance of wheeze data processed in the frequency domain is better as compared to the data processed in the time domain. A high performance accuracy with 97% is obtained comparing to an accuracy percentage of 83.13% were only achieved by using the time domain data. Thus, this paper has successfully made a comparison between the domains proving the reliability of the frequency domain for wheeze detection.