Question

find the evaluation metrices for KNN and SVM and display the best classification

1.For the following two confusion matrix: PREDICTED CLASS 1. K Nearest Neighbor Class=Yes class=No ACTUAL CLASS Class Yes 0 53 Class-No 158 9 PREDICTED CLASS 2.Support Vector Machine Class=Yes class=No ACTUAL CLASS Class=Yes 3 46 Class=No 13 158

Answer 1

Answer: K Nearest Neighbour

<<<<<<<<<<<<<<<<<<>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>

classification gets evaluated based on its accuracy of prediction of result

so for knn

total sample size=53+0+9+158=220

correctly predicted = 211

so accuracy percentage= (total_accurate/total_sample)*100=(211/220)*100=95.91%

i.e error in accuracy while using KNN is 4.091 %

for SVM

total sample size=46+3+13+158=220

correctly predicted = 46+158=204

so accuracy percentage= (total_accurate/total_sample)*100=(204/220)*100=92.723%

i.e error in accuracy while using KNN is 7.273 %

so error rate for KNN is less so best classification in this scenario is KNN

<<<<<<<<<<<<<<<<<<<<<>>>>>>>>>>>>>>>>>>>>>>>>>

note: if u want to evaluate above algo based on different criterion let me know

happy learning and please upvote