论文信息 - A nearest neighbour rule with class membership (NNRC) for modelling problems

A nearest neighbour rule with class membership (NNRC) for modelling problems

The nearest neighbour rule (NNR) has been used widely to determine a bound on the performance of classifiers. It has been shown that the error rate of the nearest neighbour classifier bounds the optimal Bayes error rate by a factor of at most two. We present NNRC, a nearest neighbour rule with class membership, to model the multiple fault conditions on a test rig. The NNR rule can be used only for classification problems. Hence we extend the NNR with NNRC to allow the use of continuous class labels as well.

A.J. Hoffman | N.T. van der Merwe | C. Stander | S.P. Heyns

[1] Christopher M. Bishop,et al. Neural networks for pattern recognition , 1995 .

[2] Asoke K. Nandi,et al. CYCLOSTATIONARITY IN ROTATING MACHINE VIBRATIONS , 1998 .

[3] Brian D. Ripley,et al. Pattern Recognition and Neural Networks , 1996 .

[4] Y.-T. Su,et al. On initial fault detection of a tapered roller bearing: Frequency domain analysis , 1992 .

[5] A.J. Hoffman,et al. Developing an efficient cross validation strategy to determine classifier performance (CVCP) , 2001, IJCNN'01. International Joint Conference on Neural Networks. Proceedings (Cat. No.01CH37222).

[6] Ibrahim Esat,et al. ARTIFICIAL NEURAL NETWORK BASED FAULT DIAGNOSTICS OF ROTATING MACHINERY USING WAVELET TRANSFORMS AS A PREPROCESSOR , 1997 .