Table 4 Classification accuracy using ALL/AML leukemia data. PA is the prediction accuracy.
From: A Java-based tool for the design of classification microarrays
| Â | Number of clusters of probes, Îş | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Â | 2 | 5 | 10 | 20 | 30 | 40 | ||||||
Number of top-ranked probes | PA (%) | No. of probes | PA (%) | No. of probes | PA (%) | No. of probes | PA (%) | No. of probes | PA (%) | No. of probes | PA (%) | No. of probes |
50 | 94.44 | 2 | 94.44 | 5 | 97.22 | 10 | 97.22 | 20 | 97.22 | 30 | 97.22 | 39 |
100 | 88.89 | 2 | 95.83 | 5 | 94.44 | 10 | 97.22 | 20 | 97.22 | 30 | 97.22 | 40 |
150 | 83.33 | 2 | 95.83 | 5 | 97.22 | 10 | 97.22 | 20 | 97.22 | 30 | 97.22 | 39 |
200 | 79.17 | 2 | 80.56 | 5 | 97.22 | 10 | 97.22 | 20 | 97.22 | 30 | 97.22 | 39 |
250 | 79.17 | 2 | 79.17 | 5 | 97.22 | 10 | 97.22 | 20 | 97.22 | 30 | 97.22 | 39 |