Abstract:
This paper introduces a modified version of the k-Nearest Neighbor (kNN) classifier named Multiple Pulse k-Nearest Neighbor (MPkNN). MPkNN is developed by combining the special characteristic of the Pulse Active Width (PAW)technique with kNN. In this work, PAW and MPkNN have been used in a Malay electrocardiogram (ECG) based Class attendance system. A step by step description of generating the new Malay electrocardiogram (ECG) database is also described in this paper. A comparison of a misclassified set of results when using k-NN, with the correct results when using MPkNN is shown in the results section. Overall, PAW and MPkNN manage to correctly identify all subjects in the database during the authentication process. This study presents a novel method of reconfirming classification results using kNN for the same investigated signal.
