1

So I am looking for a way to use a dataset with Support Vector learning where I have all the data either numeric or easily vectorizable.

Only one challenge I'm facing is with IP addresses in the Dataset.

I found a really good answer here for IPv4 addresses here. Email and IP String preprocessing for classification task

I could probably adapt the same technique for IPv6 addresses too.

But is there any preferred technique or can someone suggest some good technique for vectorizing the IPv6 Addresses for SVM?

Community
  • 1
Keyur Golani
  • 111
  • 3
  • The first thing that comes to my mind is to split each IP4 into four variables one per each digit group. – Tim Sep 29 '16 at 08:28
  • Yes. That's what the link that I posted says. But for IPv6 this approach would be a little complex. Want to know if there could be something that is more significant towards IPv6 as well as simpler to implement and explain someone. – Keyur Golani Sep 29 '16 at 08:32

0 Answers0