0

I am working on Multi-Class classification problem, where I have designed

SVM classifier

for

31 different classes

After the classification process I got the predicted labels now I want to calculate the accuracy rate for each distinct class. Although I have calculated the overall accuracy of model but I want more detail.

Can some one provide me Matlab code for this purpose or guide me to calculate the desired result.

shin
  • 31,901
  • 69
  • 184
  • 271
  • You could subset your matrix of predictions into each class, then compute (# correct classifications) / (# classifications). Without knowing the structure of your data (are classes codes as numerics vs strings, etc.) it's hard to give a complete answer. – DMR Jan 16 '17 at 04:37
  • Please provide more detail. Specifically, show us your code you used to perform classification with 31 distinct classes. Once we see the structure of your code, we can then provide an answer using your code as a base in order to make our answers easily streamlined into your code. – rayryeng Jan 16 '17 at 06:44

0 Answers0