Wrapper based gene selection methods tend to obtain better classification accuracy than filter methods, while it is much more time consuming. Accelerating this process without degrading the high accuracy is of great value for researchers to better analyze gene expression profiles. In this paper, we explore to reduce the time complexity of wrapper based gene selection method with K-Nearest-Neighbor (KNN) classifier embedded. Instead of taking KNN as a black box, we incrementally construct and maintain a classifier distance matrix to speed up the gene selection process. Experiments on six publicly available microarrays were first conducted to show the effectiveness of incremental wrapper based gene selection method with KNN. Then, to demonstrate the performance gain in time cost reduction, we analyzed the time complexity and experimentally evaluated it. Both theoretical analysis and experimental results prove that the proposed approach greatly accelerates the gene selection process without degrading the classification accuracy.
[1]
Aixia Guo,et al.
Gene Selection for Cancer Classification using Support Vector Machines
,
2014
.
[2]
Marko Robnik-Sikonja,et al.
Theoretical and Empirical Analysis of ReliefF and RReliefF
,
2003,
Machine Learning.
[3]
Jesús S. Aguilar-Ruiz,et al.
Incremental wrapper-based gene selection from microarray data for cancer classification
,
2006,
Pattern Recognit..
[4]
Huan Liu,et al.
Consistency-based search in feature selection
,
2003,
Artif. Intell..
[5]
José M. Puerta,et al.
On incremental wrapper-based attribute selection: experimental analysis of the relevance criteria
,
2008
.