Calculate miss classification percentage

Dear All;

if i need to calcuate misclassification percentage for each class, usually i looked to confusion matrix but i have a confusion which percentage represent the misclassification rate , is it the predication percentage or ture percentage ? i think it is the prediction percentage becuase it tell how the model pefrom in classifcation.

i know that there are many terms used such true positive, true negatrive ...etc but simply speaking if i want to tell what is percentage of correct classification in simple words , which one shall i use

Answer 1

[KALYAN ACHARJYA]

is it the predication percentage or ture percentage ?

As per my undestanding it is prediction percentage based on your proposed model, the true percentage is depends on ground truth data (gold data) recomended/verified by respective data experts.

In one terms, you can look for sensitivity (true positive rate) and specificity (true negative rate) or accuracy or PPV etc.

For detail, please look at Tom paper.