I am using Transfer Learning for Classification of my Dataset.

How to calculate Classification accuracy of each class?

I am using Transfer Learning for Classification of my Dataset.

How to calculate Classification accuracy of each class?

1 Like

I would just have an array `correct`

filled with zeros and size of number of total classes. Then I would classify data point `j`

, if it matches the target `label[j]`

then just increment the array with that class index, `correct[j] += 1`

. If `num_class`

is an array that contains the number of points for each class `c`

, then the accuarcy is `correct[c] / num_class[c]`

.

Answer given by @ptrblck Thanks a lot!

```
nb_classes = 9
confusion_matrix = torch.zeros(nb_classes, nb_classes)
with torch.no_grad():
for i, (inputs, classes) in enumerate(dataloaders['val']):
inputs = inputs.to(device)
classes = classes.to(device)
outputs = model_ft(inputs)
_, preds = torch.max(outputs, 1)
for t, p in zip(classes.view(-1), preds.view(-1)):
confusion_matrix[t.long(), p.long()] += 1
print(confusion_matrix)
```

To get the per-class accuracy:

```
print(confusion_matrix.diag()/confusion_matrix.sum(1))
```

2 Likes

Yes, you should calculate the accuracy on your test images. I also suggest you create stratified 5 or 10 fold experiment.

Kindly help me to make confusion matrix for the following code, where I am calculating accuracy per class for 101 class labels

video_pred = [np.argmax(x[0]) for x in output]

```
video_labels = [x[1] for x in output]
print('Accuracy {:.02f}% ({})'.format(
float(np.sum(np.array(video_pred) == np.array(video_labels))) / len(video_pred) * 100.0,
len(video_pred)))
#Accuracy per Class
for i in range(num_class):
indicies_correct = np.where(np.array(video_pred) == np.array(video_labels))
class_total = np.sum(np.array(video_labels) == i)
class_correct = np.sum(np.array(video_pred)[indicies_correct ] == i)
print('Class {}: Accuracy {:.02f}%'.format(i, class_correct / class_total * 100))
confusion_matrix = torch.zeros(num_class, num_class)
data_gen = enumerate(data_loader)
#with torch.no_grad():
for i, (data, label) in data_gen:
data = data.to(device)
label = label.to(device)
outputs = model_ft(data)
_, preds = torch.max(outputs, 1)
for t, p in zip(label.view(-1), preds.view(-1)):
print(t, p)
confusion_matrix[t.long(), p.long()] += 1
print(confusion_matrix)
print(confusion_matrix.diag()/confusion_matrix.sum(1))
```