locked
clustering algorithm for anomaly detection RRS feed

  • Question

  • Hi all,

    I have a problem in running clustering algorithm for anomaly detection in credit card fraud detection.I have one cube contain two table"card" and "Dim Date Time". also,  I have two measure sum amount and sum count of each transactions. I need these two measures in one day for each credit card number for my input model. But i dont know it is better use a Day Attribute in my "Dim Date Time" table in my cube such as a nested table, or create a daily amount and daily Count "Calculated member" at first and then using these two measures to inpute my model?????? 
    Please guid me

    Tuesday, May 5, 2009 6:45 AM

All replies

  • hi, me working on data mining in credit card fraud detection... my aim is to layout unauthorized transaction with algo's approach...kindly if anyone can help me regarding algo's approach...thanx
    Tuesday, June 9, 2009 5:11 PM