Multi-class classification using kernel density estimation on K-nearest

A fast and accurate multi-class classification method based on the conventional kernel density estimation (KDE) and K-nearest neighbour (KNN) techniques is proposed. This method estimates the cumulative probabilities of the test sample on its KNNs which may belong to different classes, then selects the maximum weighted class as the classification result. Experiments are carried out to diagnose multiple parametric faults in an analogue circuit, and the classification performances of the proposed method as well as KNN, KDE and support vector machine are compared with each other in detail. The results show that the proposed method is generally better than the other methods not only in classification accuracy but also in test speed, and is promising for practical use.