The accompanying files for my Teaching Your Machine To Find Fraudsters talk. 

clickgen2.php - generates click data to fraud2.csv
paygen.php - generated payment data to paydata.csv and paytest.csv

fraudclicks.csv - the data from the talk, with a fraud introduced at day 201
average.php - perform click detection based on the moving average + stddev
sickavail.php - perform click detection based on the sickness availability model


fraudknn.php - K Nearest Neighbour based learner for small built in fraud set
testknn.php - Nearest neighbour search, using Xapian

paydata.csv - 10000 transactions, mix of fraud/not fraud for training
paytest.csv - 1000 of the same data, generated separately for testing
learn.php - SVM learner and self evaluation on paydata data
test.php - SVM test on paytest data

learn.php and test.php require the SVM extension, available at pear.phpir.com
fraudknn.php and testknn.php require the Xapian extension and libraries, generally available through package management (sometimes as xapian-bindings)