Figure 7
From: Support Vector Machine Implementations for Classification & Clustering
The Percentage Data Rejection vs SN+SP curves are shown for test data classification runs with a multiclass discriminator. The following criterion is used for dropping weak data: for any data point xi; if maxm{fm(xi)} ≤ Confidence Parameter, then the data point xi is dropped. For this data set using AbsDiff kernel (σ2 = 0.2) performed best, and a confidence parameter of 0.8 achieve 100% accuracy.