Minimum distance classifiers Define a mean feature vector m for each class For any object, define the distance to each mean The object belongs to the closest' class Distance defined by vector norms E Class m-X mit 人 16881
16.881 MIT Minimum Distance Classifiers • Define a mean feature vector m for each class • For any object, define the distance to each mean • The object belongs to the “closest” class • Distance defined by vector norms x1 x2 x3 m1 m3 m2 x m − x 3 m1 m2 m3 x ⋅ ⋅ ⋅ Minimum Class
Distance metrics or norms Euclidean(two)norm Manhattan metric l|=∑u Infinity norm u= max(u Euclidean Manhattan Infinity mit 人 16881
Distance Metrics or Norms • Euclidean (two) norm = ∑ i i 2 2 u u • Manhattan metric u = ∑ ui i • Infinity norm u ∞ = max ( ui ) Euclidean Manhattan Infinity 16.881 MIT