mRMR-ABC: A Hybrid Gene Selection Algorithm for Cancer Classification Using Microarray Gene Expression Profiling

Table 10

Comparison between mRMR-ABC and ABC algorithms classification performance when applied with the SVM classifier for Leukemia2 dataset.

Number of genes	Classification accuracy
	mRMR-ABC			ABC
	Best	Mean	Worst	Best	Mean	Worst
2	84.72%	84.72%	84.72%	84.72%	84.72%	84.72%
3	87.5%	86.11%	84.72%	86.11%	85.23%	84.72%
4	90.27%	87.5%	84.72%	87.5%	86.11%	84.72%
5	90.27%	88.88%	86.11%	87.5%	86.45%	84.72%
6	94.44%	90.27%	87.5%	90.27%	88.88%	86.11%
7	93.05%	89.49%	88.88%	90.27%	89.22%	86.11%
8	94.44%	91.66%	87.5%	91.66%	90.27%	88.88%
9	94.44%	92.38%	87.5%	93.05%	91.46%	88.88%
10	95.83%	91.66%	88.88%	93.05%	91.98%	88.88%
15	98.61%	94.44%	91.66%	94.44%	92.78%	90.27%
18	98.61%	95.67%	91.66%	95.83%	92.99%	90.27%
20	100%	96.12%	95.83%	97.22%	93.15%	91.66%