I need the help how to calculate precision and recall from this confusion matrix (attached as image)?
As per the definition of precision and recall.
- Precision = TP / (TP+FP)
- Recall = TP / (TP+FN)
The diagonal values are considered as Recall in this confusion matrix, but
how the precision values are calculated?
how the accuracy is calculated (86.5)?
Dear Asif Nawaz the both definitions that you mentioned work completely when you have 2 classes or 2 by 2 matrix of actual and predicted sets..while for 2 or more classes a bit (like shows in your .PNG) change happens on the definitions you have to find the accuracy, precision,...,etc according to select specific diagonal element in the matrix..this image below may helps you
check this link below to get some other answers relate to your question
https://www.researchgate.net/post/How_do_you_measure_specificity_and_sensitivity_in_a_multiple_class_classification_problem
Two ways:
- get the precision and recall for each class and average - get the precision and recall for each class, and weight by the number of instances of each class. That will give you the weighted precision and recall.
weighted = (p1*s1 + p2*s2 + p3*s3 + p4*s4+p5*s5 + p6*s6)/(s1 + s2 + s3 + s4 + s5 + s6)
Where p1,p2,..p6 are precisions of class walk, bike,..plane respectively and s1, s2,...s6 are segment number or instance of each class.
- Badges
- Science topic
- Similar topics
- Computer Science
- Images