# Mapping of feature names from legacy to new

**URL:** <https://forum.numer.ai/t/mapping-of-feature-names-from-legacy-to-new/4661>\
**Category:** Tournament\
**Created:** [December 22, 2021, 2:17am UTC](https://forum.numer.ai/t/mapping-of-feature-names-from-legacy-to-new/4661 "2021-12-22T02:17:31Z")\
**Posts on this page:** 5\
**Page:** 1

<div class="post-metadata">

**Author:** ![mic](https://yyz1.discourse-cdn.com/flex009/user_avatar/forum.numer.ai/mic/32/2949_2.png) [@mic](https://forum.numer.ai/u/mic)\
**Post date:** [December 22, 2021, 2:17am UTC](https://forum.numer.ai/t/mapping-of-feature-names-from-legacy-to-new/4661/1 "2021-12-22T02:17:31Z")

</div>

Here is an unofficial mapping:

> <https://github.com/mizimno/numerai/blob/main/feature_mapping.csv>

Numerai said that the new features aren’t exactly the same as the old features, so these are the closest matches. Features that don’t exist in the new data set are left blank.

The mapping is reverse engineered from the datasets themselves. I have no knowledge of the feature engineering Numerai did to create the features.

---

<div class="post-metadata">

**Author:** ![ml\_is\_lyf](https://avatars.discourse-cdn.com/v4/letter/m/e56c9b/32.png) [@ml\_is\_lyf](https://forum.numer.ai/u/ml_is_lyf)\
**Post date:** [December 22, 2021, 7:05pm UTC](https://forum.numer.ai/t/mapping-of-feature-names-from-legacy-to-new/4661/2 "2021-12-22T19:05:34Z")

</div>

This is great, thanks! How did you determine this?

---

<div class="post-metadata">

**Author:** ![mic](https://yyz1.discourse-cdn.com/flex009/user_avatar/forum.numer.ai/mic/32/2949_2.png) [@mic](https://forum.numer.ai/u/mic)\
**Post date:** [December 23, 2021, 6:51am UTC](https://forum.numer.ai/t/mapping-of-feature-names-from-legacy-to-new/4661/3 "2021-12-23T06:51:53Z")

</div>

The pairs were selected as those with maximum average correlation (averaged across eras) and a tie break if a feature was selected in multiple pairs. From memory, I think only one tie break was needed, for `charisma72`, that resulted in `feature_acerb_venusian_piety` being paired with its second choice of `charisma7` instead, which had nearly the same correlation on both fronts.

For validation I sampled a few to check it looked right, but you might want to check a pair or two yourself before you go to town with it.

---

<div class="post-metadata">

**Author:** ![mic](https://yyz1.discourse-cdn.com/flex009/user_avatar/forum.numer.ai/mic/32/2949_2.png) [@mic](https://forum.numer.ai/u/mic)\
**Post date:** [April 11, 2022, 1:47am UTC](https://forum.numer.ai/t/mapping-of-feature-names-from-legacy-to-new/4661/4 "2022-04-11T01:47:00Z")

</div>

Updated with mapping for v3 to v4.

> **[GitHub - miciasto/numerai](https://github.com/miciasto/numerai)**
>
> Contribute to miciasto/numerai development by creating an account on GitHub.

---

<div class="post-metadata">

**Author:** ![mic](https://yyz1.discourse-cdn.com/flex009/user_avatar/forum.numer.ai/mic/32/2949_2.png) [@mic](https://forum.numer.ai/u/mic)\
**Post date:** [April 11, 2022, 3:12am UTC](https://forum.numer.ai/t/mapping-of-feature-names-from-legacy-to-new/4661/5 "2022-04-11T03:12:48Z")

</div>

sorry, just corrected some duplicated mappings, please re-download if using