Combined Gene Selection Methods for Microarray Data Analysis
- Cite this paper as:
- Hu H., Li J., Wang H., Daggard G. (2006) Combined Gene Selection Methods for Microarray Data Analysis. In: Gabrys B., Howlett R.J., Jain L.C. (eds) Knowledge-Based Intelligent Information and Engineering Systems. KES 2006. Lecture Notes in Computer Science, vol 4251. Springer, Berlin, Heidelberg
In recent years, the rapid development of DNA Microarray technology has made it possible for scientists to monitor the expression level of thousands of genes in a single experiment. As a new technology, Microarray data presents some fresh challenges to scientists since Microarray data contains a large number of genes (around tens thousands) with a small number of samples (around hundreds). Both filter and wrapper gene selection methods aim to select the most informative genes among the massive data in order to reduce the size of the expression database. Gene selection methods are used in both data preprocessing and classification stages. We have conducted some experiments on different existing gene selection methods to preprocess Microarray data for classification by benchmark algorithms SVMs and C4.5. The study suggests that the combination of filter and wrapper methods in general improve the accuracy performance of gene expression Microarray data classification. The study also indicates that not all filter gene selection methods help improve the performance of classification. The experimental results show that among tested gene selection methods, Correlation Coefficient is the best gene selection method for improving the classification accuracy on both SVMs and C4.5 classification algorithms.
Unable to display preview. Download preview PDF.