# Feature Engineering and Labeling New Variables in the Youth Mental Health Narratives

**URL:** <https://community.drivendata.org/t/feature-engineering-and-labeling-new-variables-in-the-youth-mental-health-narratives/10869>\
**Category:** Youth Mental Health: Automated Abstraction\
**Created:** [October 22, 2024, 3:21pm UTC](https://community.drivendata.org/t/feature-engineering-and-labeling-new-variables-in-the-youth-mental-health-narratives/10869 "2024-10-22T15:21:38Z")\
**Posts on this page:** 3\
**Page:** 1

<div class="post-metadata">

**Author:** ![InsightHacker](https://avatars.discourse-cdn.com/v4/letter/i/43a26b/32.png) [@InsightHacker](https://community.drivendata.org/u/InsightHacker)\
**Post date:** [October 22, 2024, 3:21pm UTC](https://community.drivendata.org/t/feature-engineering-and-labeling-new-variables-in-the-youth-mental-health-narratives/10869/1 "2024-10-22T15:21:38Z")

</div>

I am working on the Youth Mental Health Narratives project and have a few questions regarding feature engineering. Specifically, can we create new variables and add them as labels or targets in the dataset? I’m also unsure about the best practices for generating and incorporating these new variables into the model.

I’m seeking advice from the community on how to approach this, and I would greatly appreciate any insights or examples from those who have worked on similar challenges.

---

<div class="post-metadata">

**Author:** ![InsightHacker](https://avatars.discourse-cdn.com/v4/letter/i/43a26b/32.png) [@InsightHacker](https://community.drivendata.org/u/InsightHacker)\
**Post date:** [October 24, 2024, 11:31am UTC](https://community.drivendata.org/t/feature-engineering-and-labeling-new-variables-in-the-youth-mental-health-narratives/10869/2 "2024-10-24T11:31:10Z")

</div>

@kwetstone Can you support me?

---

<div class="post-metadata">

**Author:** ![ishashah](https://avatars.discourse-cdn.com/v4/letter/i/f475e1/32.png) [@ishashah](https://community.drivendata.org/u/ishashah)\
**Post date:** [October 25, 2024, 3:19pm UTC](https://community.drivendata.org/t/feature-engineering-and-labeling-new-variables-in-the-youth-mental-health-narratives/10869/3 "2024-10-25T15:19:04Z")

</div>

Hi @InsightHacker , yes, you are welcome to derive new variables to use for the task of [making predictions for the fields of interest](https://www.drivendata.org/competitions/295/cdc-automated-abstraction/page/917/#labels) for this challenge.

Our About page for this challenge contains [links to a few research papers that might be helpful in working with this dataset](https://www.drivendata.org/competitions/295/cdc-automated-abstraction/page/916/#research-using-nvdrs).