Research Article
Improved Instance Selection Methods for Support Vector Machine Speed Optimization
Table 1
Result for technique proposed by Panda et al. [
1].
| Number of instances in subset | | Average prediction accuracy | FP (%) | FN (%) | Time (s) |
| 50 | 50 | 98.95 | 1.20 | 0.00 | 30.81 | 100 | 50 | 99.90 | 0.11 | 0.00 | 40.14 | 150 | 50 | 98.70 | 1.49 | 0.00 | 29.59 | 200 | 50 | 99.08 | 1.06 | 0.00 | 33.31 | 250 | 50 | 99.63 | 0.43 | 0.00 | 39.26 | 300 | 50 | 99.78 | 0.26 | 0.00 | 44.16 | 350 | 50 | 99.78 | 0.26 | 0.00 | 54.76 | 400 | 50 | 99.78 | 0.26 | 0.00 | 57.14 |
|
|
Key: FP: false positive and FN: false negative.
|