The goal of the Dev-o-thon competition is to develop a lead scoring algorithm that can predict the probability of a customer converting into a buyer based on various features such as demographics, behavior, preferences, etc. The Benchmark Values for Submission in the Competition include following : 

| Score Matrix | Expected Score |
| --- | --- |
| F1 Score | 50% |
| Precision %'age | 66% |
| Recall %'age | 70% |


## Submission Format

The submission file should have a header and contain two columns:

- `id`: the unique identifier of each customer
- `score`: the predicted label of conversion (0 or 1)

For example:
| ID | Status |
| --- | --- |
| 985asdf5 | 0.0 |


## Evaluation Metric

The submissions will be evaluated on the **F1 score with micro averaging**. The F1 score is the harmonic mean of precision and recall, and it ranges from 0 (worst) to 1 (best). Precision is the ratio of true positives to all predicted positives, and recall is the ratio of true positives to all actual positives.

The F1 score with micro averaging can be calculated as follows:

```
F1_micro = 2 * (TP / (TP + FP)) * (TP / (TP + FN)) / ((TP / (TP + FP)) + (TP / (TP + FN)))
```

Where TP is the total number of true positives across all classes, FP is the total number of false positives across all classes, and FN is the total number of false negatives across all classes.

The F1 score with micro averaging gives equal weight to each class and each sample, and it is suitable for imbalanced datasets.

The higher the F1 score, the better the performance of the lead scoring algorithm.

## Leaderboard

The leaderboard will show the public F1 score of each submission based on a subset of the test data. The final ranking will be determined by the private F1 score based on the remaining test data, which will be revealed after the competition ends.