Python Forum

Hey guys,

I have a simple Random Forest Regression Model which is supposed to predict the Number of Likes of an Instagram Picture based on given features.
On some pictures the results are amazing, on others not so much. Hence I want to take a look at the content of the pictures where the prediction is bad. To be more specific I want a graph which shows the true values of the test set vs the predicted values which have an accuracy lower than 80% and moreover list the row of the image in the test data frame.
Is something like this possible? thanks in advance!

Yes, but a bit more manual than usual.
For your test set, loop through using your prediction function to get a predicted value, compare to the real value, and if off by more than 20% then note that row in your dataframe