* Test dataset statistics

| Feature         | Expression                                                                                 |
|-----------------+--------------------------------------------------------------------------------------------|
| card            | RANDBETWEEN(0,100)                                                                         |
| amount          | RANDBETWEEN(0,1000)                                                                        |
| event_timestamp | RANDOM                                                                                     |
| is_fraud_label  | 400<amount<700 & cat1_string="C" & num1_float < 700                                        |
| uuid            | UNIQUE                                                                                     |
| cat1_generator  | RANDBETWEEN(0,3)                                                                           |
| cat2_generator  | RANDBETWEEN(0,2)                                                                           |
| cat3_generator  | RANDBETWEEN(0,5)                                                                           |
| num1_float      | RANDBETWEEN(1,1000)+0.4                                                                    |
| num2_double     | RANDBETWEEN(200000,40000000)-0.3                                                           |
| num3_int        | RANDBETWEEN(10000,12000)                                                                   |
| cat1_string     | IFS(F2=0,"aaa",F2=1,"b",F2=2,"C",F2=3,"dDd")                                               |
| cat2_string     | IFS(G2=0,"aaacat2",G2=1,"bcat2",G2=2,"Ccat2",G2=3,"dDdcat2")                               |
| cat3_string     | IFS(H2=0,"aaacat3",H2=1,"bcat3",H2=2,"Ccat3",H2=3,"dDdcat3",H2=4,"sdofij",H2=5,"blahblah") |

| Dataset          |      N | N_fraud | Fraud rate (%) |
|------------------+--------+---------+----------------|
| train.csv        | 184951 |    9792 |            5.3 |
| in_train_val.csv |  50000 |    2719 |            5.5 |
| val.csv          |  50000 |    2569 |            5.1 |