How to Analyz Clustering Output

Replies

• 

  Manish Goyal

  hi steward
  
  Frankly saying i have no idea about this rapid miner but i am interested in helping you in this
  
  If you can explain what is this rapid miner all about in brief (means what is the output and input?) then may be i can help you.
• 

  Reya

  I have learnt about Rapid Miner tool in data mining..Can u explain me about Botnet detection tool?I have no idea about that.
• 

  steward

  Hi
  
  First, Rapidminer is a Datamining software that consists of Classification and Clustering Algorithms such as Support Vector machine(SVM) or Xmeans and so on.
  
  The traffic that I m going to detect the botnet involved, has lot of records and flows inside.Therefore for the first step I need to cluster those heavy records to work easier and get involved with organized records.
  
  Bot is a kind of malware that after infecting your computer, it can connect back to its bot master(Hacker) or Command and Control server(C&C).
  So botnet is a collection of computers that have been infected by the hacker to be misused for illegal actions such DDOS, Spamming....
  
  For my work I wanna to detect the flows that are connecting with their C&C based on a particular Framework.
  
  I wanna use Pcap dataset for my work...And Inside the pacp file is all the packet information (src add, src mac, des add, des port, src port,,and so on).
• 

  Manish Goyal

  ok so you want to know how to use the data stored in pacp file?
  
  am i right?
• 

  steward

  goyal420

  ok so you want to know how to use the data stored in pacp file?
  
  am i right?

  But for this thread ..no ...I want to input Dataminer with Pcap file and as i explained to cluster with Xmeans algo....
  
  By the way, I m using java library Jpcap and Jnetpcap (I do not know which of them is better).but thank you for asking. because I am searching any java flow analysis tools that can calculate some statistical analysis on flow such as flow size or average packet size of a flow and so on. do you know any one?
• 

  steward

  No have been working on clustering?