|
The relationship between accuracy (%) and the filter threshold The cross validation accuracies with Naive Bayes (NB) and Nearest Neighbor (NN) Methods are displayed with different filter cutoffs. The percentages of the features (gene pairs) used are listed as well. Since the number of features used are slightly different for different test samples, ranges are shown. |
||||||
| Filter Threshold |
No Filter |
0.02 |
0.04 |
0.05 |
0.06 |
0.08 |
|
|
||||||
| NB |
63 |
66 |
68 |
69 |
70 |
68 |
| NN |
70 |
73 |
75 |
77 |
75 |
74 |
| Features used (%) |
100 |
41–43 |
12–13 |
4.9–5.9 |
1.7–2.1 |
0.12–0.17 |
Wang and Huang BMC Bioinformatics 2004 5:121 doi:10.1186/1471-2105-5-121 |
||||||