CODEWITHSUNDEEP
svmnovelty-detection
Bubba
Bubba

Reputation: 105

Novelty detection alternatives to one class SVM

I have a case where I need to detect anomalies in a dataset. I do not have a labeled trainingset. The problem is not entirely unsupervised either since I know part of the data contains no anomalies, and I know which part that is. This is called novelty detection, as far as I'm aware. One class SVM would work here but I'd like to compare it to alternative methods.

So far I've only found methods that are entirely unsupervised or entirely supervised, so no valid alternatives to one class SVM for my novelty detection case. Does anyone know of methods that would be valid alternatives (preferably available in Python)?

Upvotes: 0

Views: 781

Answers (1)

tolgayilmaz
tolgayilmaz

Reputation: 4195

There is a textblob module and it has a PositiveNaiveBayesClassifier which is more practical to code than OneClassSVM. Have a look at that, It will suffice your needs.

Upvotes: 0

Related Questions