dm.cs.tu-dortmund.de/en/mlbits/outlier-one-class-svm/
One-Class Classification – Lecture Notes
t to}\quad & \lVert x_i - a \rVert ^2 \leq R^2 + \xi _i & & \text{for } i=1,2,\dots ,N \\ & \xi _i \geq 0 & & \text{for } i=1,2,\dots ,N \end{align*}\]
one-class classification, positive examples only [...] x_i) \\ \text{subject to}\quad & 0 \leq \lambda _i \leq C & & \text{for } i=1,2,\dots ,N \\ & \sum \nolimits _i \lambda _i = 1 \end{align*}\]
which also can be solved by existing solvers.
The center (with [...] kernel
related to density estimation (level sets)
SVDD Dual Problem
With the Lagrange approach we obtain 1 the following dual problem:
\[\begin{align*} \text{minimize }\quad & \sum \nolimits _{i,j} \lambda _i\lambda …