Can someone tell me if my machine learning model is overfitting or not ?
I hope you are doing well , I m working on a machine learning project and I just plotted the confusion matrix
9 Replies
Can someone tell me if my model is overfitting or not ?
the model performs well for some classes but doesn't with others and that is a sign of overfitting
What are the classes that have overfitting if you can help ?
'e' 'd' 'm' 'n' 'o' 'p' 'q' 'r's' 't' 'u' 'v' 'w' 'x' 'y' 'z' 'you' 'bad 'fine' 'hungry' 'sorry' 'thankyou' 'deaf' 'me' 'please' 'goodbye', 'no' , the diagonal values are relatively low
But these have a value of 98% , how they can be overfitted ?
the diagonal values (representing correct classifications) are relatively low compared to the off-diagonal values (misclassifications). This suggests that the model struggles to accurately classify instances belonging to these classes.
@anniekeerdekens
It is difficult to determine only on the confusion matrix if your model is overfitting. What data is the confusion matrix representing? Only your validation data? Is the validation data unseen? How did you split the data? Is the data balanced? I suggest you first have a look at your loss curves. Check out this article: https://towardsdatascience.com/learning-curve-to-identify-overfitting-underfitting-problems-133177f38df5.