Will Koehrsen
1 min readAug 14, 2018

--

Feature engineering is crucial and you’re right that if we have better features, we don’t need such overly complex models. Moreover, creating deep features through DFS can let us extract useful information about real-world problems such as which variables are most important to monitor.

Thanks for the resource. That’s an interesting approach to feature engineering and I’m eager to try it out!

--

--

Will Koehrsen
Will Koehrsen

Written by Will Koehrsen

Senior Machine Learning Engineer at Cortex Sustainability Intelligence

No responses yet