interpret the performance of my model from confusion matrix

Question

how I can interpret the performance of my model from this matrix

Answer 1

Its always better to evaluate through precision, recall and f-score. precision is defined as True positive/ True positives + False Positives and precision is True positive/ True positives + False Negatives. F-score is an mean of both precision and recall.

There is a way clearer than this. You can get the precision, recall and f-score from this function using the classification_report() from sklearn:

from sklearn.metrics import classification_report
print(classification_report(y_test, y_predicted))

Please check your y_predicted that its the right classes values before you pass it to this function you may need to use np.argmax()

An example of what the function will return something like this:

          precision    recall   f-score
class1       0.79       0.98      0.80
class2       0.89       0.79      0.82

You can find more about this function here