Excellent! Congrats!
As a hint for a potential next step. Also:
can be useful. Especially the latter can be very helpful since it visualises the whole distribution of your training and e.g validation set and you can easily compare it against your business requirements (e.g. when predicting a pressure in a regression task, your business requirement might be:
less than 110 mbar pressure error allowed.
)
The residuals should be normally distributed and they should not show any patterns or correlations with your features.
Otherwise it might be worth to exploit this and incorporate this knowledge into your features.
Best regards
Christian
Okay. Thanks.
But since the model features are currently linear, I would love to learn the ways to create polynomial features from the datasets and train my model on that first.
I believe a step at a time would help me understand the rudimentary deeply.
1 Like
… the model features are currently linear
As far as I have understood your first post, you have used a linear model.
In this case it would be helpful to model the non-linearity already in your features with feature engineering techniques. I think this is also what you are saying with polynomial features, right?
Also here the residual analysis might be a helpful tool to check how successful you were also with respect to systematics in your residuum as mentioned in the previous post. If you see systematics you can design your features w/ mathematical operations (e.g. with polynomial approaches as you mentioned before).
Anyway - good luck! Feel free to post an update or ask any question if something is open or unclear.
Best regards
Christian
Yes.
Okay, I will have a look at the residual analysis. Thanks.
1 Like
Ok. Here you can find an exemplary implementation which might be helpful.
Good luck and have fun! 
Best regards
Christian
