Feature selection for high dimensional microarray gene expression data via weighted signal to noise ratio

Abstract

Feature selection in high dimensional gene expression datasets not only reduces the dimension of the data, but also the execution time and computational cost of the underlying classifier. The current study introduces a novel feature selection method called weighted signal to noise ratio (WSNR) by exploiting the weights of features based on support vectors and signal to noise ratio, with an objective to identify the most informative genes in high dimensional classification problems. The combination of two state-of-the-art procedures enables the extration of the most informative genes. The corresponding weights of these procedures are then multiplied and arranged in decreasing order. Larger weight of a feature indicates its discriminatory power in classifying the tissue samples to their true classes. The current method is validated on eight gene expression datasets. Moreover, results of the proposed method (WSNR) are also compared with four well known feature selection methods. We found that the (WSNR) outperform the other competing methods on 6 out of 8 datasets. Box-plots and Bar-plots of the results of the proposed method and all the other methods are also constructed. The proposed method is further assessed on simulated data. Simulation analysis reveal that (WSNR) outperforms all the other methods included in the study.

Fig 1. Flowchart of the proposed method.

Fig 2. Bar-plots of error rates of the proposed and the other classical methods on various subsets for Leukemia dataset.

Fig 3. Bar-plots of error rates of the proposed and the other classical methods on various subsets of genes for Colon dataset.

Fig 4. Bar-plots of error rates of the proposed and the other classical methods on various subsets of genes for Lungcancer dataset.

Fig 5. Bar-plots of error rates of the proposed and the other classical methods on various subsets of genes for Srbct dataset.

Fig 6. Bar-plots of error rates of the proposed and the other classical methods on various subsets of genes for DLBCL dataset.

Fig 7. Bar-plots of error rates of the proposed and the other classical methods on various subsets of genes for Breast dataset.

Fig 8. Bar-plots of error rates of the proposed and the other classical methods on various subsets of genes for TumorC dataset.

Fig 9. Bar-plots of error rates of the proposed and the other classical methods on various subsets of genes for Prostate dataset.

Fig 10. Box-plots of the error rates produced by random forest, using top 10 features selected by different feature selection methods for Leukemia dataset.

Fig 11. Box-plots of the error rates produced by random forest, using top 10 features selected by different feature selection methods for Colon dataset.

Fig 12. Box-plots of the error rates produced by random forest, using top 10 features selected by different feature selection methods for Lungcancer dataset.

Fig 13. Box-plots of the error rates produced by random forest, using top 10 features selected by different feature selection methods for Srbct dataset.

Fig 14. Box-plots of the error rates produced by random forest, using top 10 features selected by different feature selection methods for DLBCL dataset.

Fig 15. Box-plots of the error rates produced by random forest, using top 10 features selected by different feature selection methods for Breastcancer dataset.

Fig 16. Box-plots of the error rates produced by random forest, using top 10 features selected by different feature selection methods for TumorC dataset.

Fig 17. Box-plots of the error rates produced by random forest, using top 10 features selected by different feature selection methods for Prostate dataset.

Fig 18. Bar-plots of errors produced by different feature selection methods on simulated data having outliers, for various subsets of genes.

Fig 19. Bar-plots of errors produced by different feature selection methods on simulated data, having no outliers, for various subsets of genes.