Calculating accuracy from a multi-class confusion matrix

Report a typo

Hey there! This problem might be a bit unpredictable, but give it a go and let us know how you do!

In the context of classification metrics in machine learning, consider the following multi-class confusion matrix:

[[30, 10, 5],
 [10, 40, 5],
 [5, 5, 55]]

The confusion matrix can be represented as follows:

[[TP_class1, FN_class1, FN_class1],
 [FP_class1, TP_class2, FN_class2],
 [FP_class1, FN_class3, TP_class3]]

If we consider the first class as positive and combine the other two classes as negative, what is the accuracy of the model in percentage, calculated as (True Positives + True Negatives) / Total Predictions?

The True Negatives can be calculated as TN = TP_class2 + TP_class3.

Select one option from the list

Create a free account to access the full topic