Open needpeeling opened 2 years ago
Is there a way of getting accuracy of the predicted class?
For example: If we have 2 classes to predict (cats, dogs) but I want to classify my test file bird.wav:
c, p, p_nam = aT.file_classification(testfiles, "svm_dog_cat","svm_rbf")
Output: bird.wav: P(media/dog_samples=0.17790907874998627) P(media/cat_samples=0.8220909212500138)
bird.wav: P(media/dog_samples=0.17790907874998627) P(media/cat_samples=0.8220909212500138)
This is obviously wrong. Therefore I´m looking for an accuracy value of predicted classes. Do you have any ideas?
Is there a way of getting accuracy of the predicted class?
For example: If we have 2 classes to predict (cats, dogs) but I want to classify my test file bird.wav:
c, p, p_nam = aT.file_classification(testfiles, "svm_dog_cat","svm_rbf")
Output:
bird.wav: P(media/dog_samples=0.17790907874998627) P(media/cat_samples=0.8220909212500138)
This is obviously wrong. Therefore I´m looking for an accuracy value of predicted classes. Do you have any ideas?