2. Consider a two-class classification problem with 6 training samples as showed in the Table and...

Transcribed Image Text:

2. Consider a two-class classification problem with 6 training samples as showed in the Table and Figure below. Each training sample Sk, k = {1, 2, 3, 4, 5, 6} has two attributes (Xk, yk) and a class label C₁ € {+1,−1}. k 1 2 3 4 5 6 XK -1 0 -1 2 3 2 Ук 0 -1 -2 0 1 2 Ck -1 -1 -1 1 1 1 OG 2 1 1 y 4 -2 1 N₂ 2 D 3 (i) Consider a test sample t whose two attributes are given as (1,0). Determine the class label of the test sample t if a 3-nearest neighbor classifier using the L₁ norm (city block or Manhattan) dissimilarity measure is used to classify the test sample. (ii) Repeat question (i) if the cosine distance (i.e., 1.0 - cosine similarity) is used to perform the 3-nearest neighbor classification. (iii) Draw the decision boundary of the support vector machine that can classify the 6 training samples into two classes. Indicate all the support vectors. X 2. Consider a two-class classification problem with 6 training samples as showed in the Table and Figure below. Each training sample Sk, k = {1, 2, 3, 4, 5, 6} has two attributes (Xk, yk) and a class label C₁ € {+1,−1}. k 1 2 3 4 5 6 XK -1 0 -1 2 3 2 Ук 0 -1 -2 0 1 2 Ck -1 -1 -1 1 1 1 OG 2 1 1 y 4 -2 1 N₂ 2 D 3 (i) Consider a test sample t whose two attributes are given as (1,0). Determine the class label of the test sample t if a 3-nearest neighbor classifier using the L₁ norm (city block or Manhattan) dissimilarity measure is used to classify the test sample. (ii) Repeat question (i) if the cosine distance (i.e., 1.0 - cosine similarity) is used to perform the 3-nearest neighbor classification. (iii) Draw the decision boundary of the support vector machine that can classify the 6 training samples into two classes. Indicate all the support vectors. X

Answer rating: 100% (QA)

This question involves solving classification problems using different methods the knearest neighbor classifier with different distance metrics and support vector machine SVM i For the knearest neighb... View the full answer