# Input data normalization

**URL:** <https://community.deeplearning.ai/t/input-data-normalization/151139>\
**Category:** Improving Deep Neural Networks: Hyperparameter tun\
**Tags:** coursera-platform\
**Created:** [June 30, 2022, 6:27pm UTC](https://community.deeplearning.ai/t/input-data-normalization/151139 "2022-06-30T18:27:27Z")\
**Posts on this page:** 3\
**Page:** 1

<div class="post-metadata">

**Author:** ![IZZETTIN\_ALHALIL](https://avatars.discourse-cdn.com/v4/letter/i/2acd7d/32.png) [@IZZETTIN\_ALHALIL](https://community.deeplearning.ai/u/IZZETTIN_ALHALIL)\
**Post date:** [June 30, 2022, 6:27pm UTC](https://community.deeplearning.ai/t/input-data-normalization/151139/1 "2022-06-30T18:27:27Z")

</div>

hello every body  
is it very important to normalize my input data if it has different feature scales?  
is this process only for making the algorithm run faster?

---

<div class="post-metadata">

**Author:** ![Christian\_Simonis](https://sea1.discourse-cdn.com/dlai/user_avatar/community.deeplearning.ai/christian_simonis/32/142583_2.png) [@Christian\_Simonis](https://community.deeplearning.ai/u/Christian_Simonis)\
**Post date:** [June 30, 2022, 6:37pm UTC](https://community.deeplearning.ai/t/input-data-normalization/151139/2 "2022-06-30T18:37:09Z")

</div>

Hi there,

yes, in my experience it is an important step so that the training works effectively and all features are treated equally within your training process. Usually you want this. E.g. for gradient-based methods this also helps to reach the optimum more quickly. I recently read a nice outline here:

> **[Feature Scaling | Standardization Vs Normalization](https://www.analyticsvidhya.com/blog/2020/04/feature-scaling-machine-learning-normalization-standardization/)**
>
> Learn why to perform feature scaling on data in machine learning. Know when to use normalization over standardization and vice-versa with a Sklearn example.

Personally, I made good experience with giving it in advance some thought how the features are distributed in reality in order to chose the right method for normalising the range, e.g.

- z Transformation / standardization for normally distributed features
- min/max scaling for e.g uniform distributed features

Hope that helps!

Best  
Christian

---

<div class="post-metadata">

**Author:** ![IZZETTIN\_ALHALIL](https://avatars.discourse-cdn.com/v4/letter/i/2acd7d/32.png) [@IZZETTIN\_ALHALIL](https://community.deeplearning.ai/u/IZZETTIN_ALHALIL)\
**Post date:** [July 12, 2022, 11:45am UTC](https://community.deeplearning.ai/t/input-data-normalization/151139/3 "2022-07-12T11:45:56Z")

</div>

appreciated  
many thanks
