Optimal Separating Hyperplane
Jump to navigation
Jump to search
An Optimal Separating Hyperplane is a Hyperplane that cleanly separates the points in a dataset X into sets that contain only members of their own type, such that the Margin between the Hyperplane and any point is maximal.
- Context:
- The hyperplane is in canonical form w.r.t. the data if the minimum distance between any data point in X and the hyperplane is 1.
- Is an Optimal Canonical Separating Hyperplane when the distance from a closest vector (or Support Vector) to the hyperplane (i.e. ((w·x)+b)/ ||w||) is equal to 1.
- See: Support Vector Machine, Primal Optimization Task.
