A question that popped up while looking at the cost function videos of Week 1: What is the relationship between cost (or loss) and accuracy? I remember I heard that accuracy is more of a human-readable format. But I hope to know how accuracy is calculated. Is it what the forum can cover? Thanks. E
Hi @Eitoy, great question
Loss is the error, how far away is your predicted value from the real value in regression problems, in classification, how many bad answers can you get, while accuracy you can think about how good your model is at predicting the real value (this is a simplify answer so you can get the intuition)
This article provides more information about it Accuracy and precision - Wikipedia
But for simplicity, check the image in which accuracy is defined as the sum of true positive and true negative values divided by the full sample.
I hope this helps!
1 Like
Hi @pastorsoto ,
This is very helpful, thank you!
1 Like