Cross Validated is a question and answer site for people interested in statistics, machine learning, data analysis, data mining, and data visualization. It's 100% free, no registration required.

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

1) I wish to compare the performance of KNN and SVM on ROC chart.

a)For KNN, I obtained a linear line on ROC Chart instead of curved one.From x=0 until x=0.275, y= 0.65 and from x=0.35 until x= 0.95, y = 1.0.

b)For SVM , the ROC curve behave normally . x=0, y=0 and at x=0.95, y=0.98. Which one is the better model of the two?

share|improve this question
For $k$ nearest neighbours, the ROC can have points spaced by $\frac{1}{k}$ only. How many nearest neighbours do you use? – cbeleites Nov 22 '12 at 10:29
2) is another question, ask it separately, please. – cbeleites Nov 22 '12 at 10:29
To tell which is the better model, one would need to know the application. I.e. we cannot possibly answer that question. – cbeleites Nov 22 '12 at 10:30
@cbeleites I used only 1 nearest neighbour. Sorry I don't understand what do you mean by application. Do you mean the Software or the parameters used? I used RapidMiner and measure types: mixed measure and mixed measure :mixed euclidean. – Siga Nov 22 '12 at 12:25
1nn: what is then the continuous score you vary over the ROC? Application: the problem you want to solve, not the software. – cbeleites Nov 22 '12 at 12:59

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Browse other questions tagged or ask your own question.