OiO.lk Community platform!

Oio.lk is an excellent forum for developers, providing a wide range of resources, discussions, and support for those in the developer community. Join oio.lk today to connect with like-minded professionals, share insights, and stay updated on the latest trends and technologies in the development field.
  You need to log in or register to access the solved answers to this problem.
  • You have reached the maximum number of guest views allowed
  • Please register below to remove this limitation

Is there a way of measuring the initial shap value of two features that have been used to create a third, using a ratio?

  • Thread starter Thread starter Melvin BARBAUX
  • Start date Start date
M

Melvin BARBAUX

Guest
In the context of machine learning on python, regression or classification whatever, I sometimes do the ratio of two features to make a third feature and drop the first two.

Code:
example : living_surface_ratio = residential_building__total_living_surface / building__footprint_surface

So at the end of my data_processing_pipe I only have living_surface_ratio as a feature.

I'd like to know if there's a way of measuring the shape value of residential_building__total_living_surface and building__footprint_surface despite the fact that the features are drop and therefore no longer available.

Maybe :

Code:
shape(living_surface_ratio) = 0.3 * shape(residential_building__total_living_surface) + 0.7µbuilding__footprint_surface

I thought about training the model with the two features and then with just the ratio, but what bothers me is that I don't really provide an interpretability of the model with the feature ratio.

I've looked at the implementation of shape value but it seems complicated to modify.

If you have any ideas, I'd love to hear from you!

Thanks in advance
<p>In the context of machine learning on python, regression or classification whatever, I sometimes do the ratio of two features to make a third feature and drop the first two.</p>
<pre><code>example : living_surface_ratio = residential_building__total_living_surface / building__footprint_surface
</code></pre>
<p>So at the end of my data_processing_pipe I only have living_surface_ratio as a feature.</p>
<p>I'd like to know if there's a way of measuring the shape value of residential_building__total_living_surface and building__footprint_surface despite the fact that the features are drop and therefore no longer available.</p>
<p>Maybe :</p>
<pre><code>shape(living_surface_ratio) = 0.3 * shape(residential_building__total_living_surface) + 0.7µbuilding__footprint_surface
</code></pre>
<p>I thought about training the model with the two features and then with just the ratio, but what bothers me is that I don't really provide an interpretability of the model with the feature ratio.</p>
<p>I've looked at the implementation of shape value but it seems complicated to modify.</p>
<p>If you have any ideas, I'd love to hear from you!</p>
<p>Thanks in advance</p>
 
Top