# Eligible features which can be used in energy forecasting

**URL:** <https://community.drivendata.org/t/eligible-features-which-can-be-used-in-energy-forecasting/2188>\
**Category:** Power Laws\
**Created:** [March 13, 2018, 5:30pm UTC](https://community.drivendata.org/t/eligible-features-which-can-be-used-in-energy-forecasting/2188 "2018-03-13T17:30:43Z")\
**Posts on this page:** 6\
**Page:** 1

<div class="post-metadata">

**Author:** ![Shanka](https://avatars.discourse-cdn.com/v4/letter/s/a183cd/32.png) [@Shanka](https://community.drivendata.org/u/Shanka)\
**Post date:** [March 13, 2018, 5:30pm UTC](https://community.drivendata.org/t/eligible-features-which-can-be-used-in-energy-forecasting/2188/1 "2018-03-13T17:30:43Z")

</div>

Can my features have an increasing trend?Will it violate my eligibility criteria?  
Only year cannot be used in the calculation of any feature right?

---

<div class="post-metadata">

**Author:** ![bull](https://yyz2.discourse-cdn.com/flex028/user_avatar/community.drivendata.org/bull/32/8_2.png) [@bull](https://community.drivendata.org/u/bull)\
**Post date:** [March 13, 2018, 5:38pm UTC](https://community.drivendata.org/t/eligible-features-which-can-be-used-in-energy-forecasting/2188/2 "2018-03-13T17:38:53Z")

</div>

If you fit your increasing trend on future data, it is not eligible.

---

<div class="post-metadata">

**Author:** ![Shanka](https://avatars.discourse-cdn.com/v4/letter/s/a183cd/32.png) [@Shanka](https://community.drivendata.org/u/Shanka)\
**Post date:** [March 13, 2018, 5:43pm UTC](https://community.drivendata.org/t/eligible-features-which-can-be-used-in-energy-forecasting/2188/3 "2018-03-13T17:43:45Z")

</div>

could you explain with an example other than year?

---

<div class="post-metadata">

**Author:** ![bull](https://yyz2.discourse-cdn.com/flex028/user_avatar/community.drivendata.org/bull/32/8_2.png) [@bull](https://community.drivendata.org/u/bull)\
**Post date:** [March 13, 2018, 5:46pm UTC](https://community.drivendata.org/t/eligible-features-which-can-be-used-in-energy-forecasting/2188/4 "2018-03-13T17:46:01Z")

</div>

If you use a trend at prediction time, that trend should be calculated on the data which is in the past when you make that prediction.

---

<div class="post-metadata">

**Author:** ![LastRocky](https://avatars.discourse-cdn.com/v4/letter/l/5f9b8f/32.png) [@LastRocky](https://community.drivendata.org/u/LastRocky)\
**Post date:** [March 13, 2018, 5:47pm UTC](https://community.drivendata.org/t/eligible-features-which-can-be-used-in-energy-forecasting/2188/5 "2018-03-13T17:47:31Z")

</div>

So, features like days since the start of the training period are not allowed, right?

---

<div class="post-metadata">

**Author:** ![Shanka](https://avatars.discourse-cdn.com/v4/letter/s/a183cd/32.png) [@Shanka](https://community.drivendata.org/u/Shanka)\
**Post date:** [March 13, 2018, 5:48pm UTC](https://community.drivendata.org/t/eligible-features-which-can-be-used-in-energy-forecasting/2188/6 "2018-03-13T17:48:41Z")

</div>

This means i can use past years in my calculation of features.Suppose i have from year 2015 in test data,then i can use upto year 2014 in my feature calculations right?