A simple object recognition system encodes objects using 3-element feature vectors. 5 objects from 3 different classes are encoded as follows: Feature vector

(0.5,0.4,0.5) Class(1)

(0.6,0.2,0.1) Class(2)

(0.3,0.3,0.4) Class(3)

(0.6,0.1,0.5) Class(2)

(0.8,0.4,0.5) Class(1)

A new object of an unknown class has a feature vector (-0.0, -0.9, 0.7). Using the Euclidean distance as the similarity measure, determine the classification of the new object using: i) a nearest mean classifier ii) a nearest neighbor classifier i) a k-nearest neighbor classifier, with k=3 ANSWER: The minimal distance computed by the nearest mean classifier over classes: Answer (to 2 decimal places) The minimal distance computed by the nearest neighbor classifier over objects: Answer (to 2 decimal places) The predicted class of the k-nearest neighbor classifier: Answer