Explain ROC curve, use relevant example or explain in simple language
Question
ROC curve is used for performance measurement for the classification problem.
Explain with example and diagram
in progress
0
Statistics
55 years
6 Answers
869 views
Grand Master 0
Answers ( 6 )
In Machine Learning, performance measurement is an essential task. So when it comes to a classification problem, we can count on an AUC – ROC Curve. When we need to check or visualize the performance of the multi-class classification problem, we use the AUC (Area Under The Curve) ROC (Receiver Operating Characteristics) curve.The ROC curve is plotted with TPR against the FPR where TPR is on the y-axis and FPR is on the x-axis. TPR- true postive rate and FPR- False positive rate. ROC is curve of possibility
The curve plots 2 parameters:
1. True Positive rate: TP/TP+FN
2. False Positive rate: FP/FP+TN
It plots TPR vs FPR at different classification thresholds.
In ROC curve, higher X value indicates higher no of false positives. While higher Y value indicates higher no of false negatives.
ROC curve or the Receiver Characteristic Operation Curve is used to show the performance of our classification model at the different classification thresholds. It takes into account two parameters: the true-positive rate (TPR) and the false-positive rate (FNR).
In ROC curve, TPR is plotted against the FPR at all threshold values of our classification model. The lower the threshold we keep for our model, the higher the rate of getting true positives as well as false positives.
The higher the AUC curve, the better it is at predicting the correct classes for our model.
The diagram for the same is shown here. This image has been taken from the google developer site.
Higher the area of the ROC curve, more accuracy the model has. That is all I know about it, without googling 😛
This can start a looong conversation with interviewer. But, surely better to have something on the desk 😛
Try to explore more or go through the answers of other users
The Receiver operation characteristic ( ROC) curve is an evaluation metric for classification problem. It is a probability curve that plot the TPR against FPR at various threshold value and essentially separates the “signal” from the”noise” the area under the curve (AUC) is the measure of the ability of a classifier to distinguish between classes and is used as a summary of the ROC curve.
The higher the AUC , the better the performance of the model at distinguishing between the positive and negative classes.
when AUC = 1 then the classifier is able to perfectly distinguish between all the positive and negative class point correctly. if however the AOC had been 0, then the classifier would be predicting all negative as positive and all positive as negative. when 0.5<AUC<1 there is high chance that the classifier will be able to distinguish the positive class value from the negative class values. this is so because the classifier is able to detect more number of true positive and true negative then false negative and false positive.
when a you sequel to 0.5 then the classify is not able to distinguish between positive and negative class point meaning either the classifier is predicting random class or constant class for all the data point.
so the higher the AUC value for the classifier the better is ability to distinguish between positive and negative class.