Performance of the logistic regression model will be measure by its power to separate good and bad clients - Gini coefficient.

---

## Submission Format

For every out of sample application ID (SK_ID_CURR), submission files should contain two columns: SK_ID_CURR and estimated probability of default or its logit transformation (score).

The file should contain a header, be in cvs file with comma as a separator. First column is named SK_ID_CURR and the second one SCORE.

```
SK_ID_CURR,SCORE
1111111,0.1
```