Credit: https://medium.com/intuitionmachine/google-and-ubers-best-practices-for-deep-learning-58488a8899b6
It is a one step process after Data Preparation and before Modeling (it is predictive modeling not the other modeling ;)). From my understanding, Feature Store is focusing on storing the features that you use in your model. For instance, you build a model for a recommendation system, the features used in that model will be stored in the Feature Store.
What's the point of storing this features?
- To keep track what are the features you use?
- Which model use this features?
- Which project use this features?
- Which features always produce higher accuracy?
- Help your teammates share their features too!
Right now, I am still building the Feature Store myself using Python and store it into a database.
You can read here to know more how it is used in big tech company (Uber).
Thank you for reading my post. :) Why not you spare some time watch this music video. Hmm feeling nostalgic right?