Though process to calculate error rate for a classification algorithm with 1000 objects?

Question

I am trying to solve this question

A classification algorithm classifies 1000 objects in to one of two classes. It incorrectly classifies 13 out of 100 class 1 objects and 53 class 2 objects. (a) What is the overall error rate? (b) What proportion of those predicted to belong to class 1 are correctly classified?

Answer

a). FP + FN =13 TP + TN =153-13 =140 Overall error rate 13/153 =8.49 As far I see it, but seems something is odd, Is this correct

b.) For b section how can I take separate error rate of class 1?

Thanks

Answer 1

a) $$ER = \frac{errors}{all}=\frac{13+53}{1000} = 0.066$$ b) $$ ER1 = \frac{True1}{True1 + False2} = \frac{87}{87+53} = \frac{87}{140} = 0.62142 $$