Скачать или смотреть Anovos Feature Engineering Demo Video 5 - Producing Model Ready Features

Anovos is an open source library, enabling scalable, efficient, end to end feature engineering. It implements a systematic & procedural data pipeline with enterprise data at one end, and model-ready features at the other.

Video 5 of the series focuses on producing model ready features with Anovos, the final step of our feature engineering process. In this video we subject our candidate feature list to a set of checks to make sure they lend themselves to be used in ML processes. In Anovos, we support four distinct types of these checks – cardinality treatment, handling missing values, categorical value encoding and dimensionality reduction.

This is part of the Anovos Demonstration Video series which consists of 5 videos:
· Video 1: Anovos intro
· Video 2: Anovos EDA using Reports
· Video 3: Anovos EDA using the data Analyzer Module
· Video 4: Feature Composition in Anovos: Producing candidate feature sets
· Video 5: Pruning candidate feature sets to produce final model-ready features

While each video discusses a set of related capabilities in Anovos, they do follow a structured feature engineering workflow that we have successfully operationalized in over 300 distinct models. Thus, while you can select any one of these videos to watch out of turn, we recommend that you consume them in the order we’ve structured them, I.e., video 1 before video 2, video 2 before video3 and so on.