Date: Tuesday, April 23, 2019

Time: 11:00 AM Eastern Daylight Time

Duration: 1 hour

World-class feature engineering -- generated automatically with minimal data science expertise

Automated feature engineering is one of the most exciting new capabilities released for machine learning platforms.

Manual feature engineering is tedious, yet it's one of the most important tasks in machine learning and data science. Generating features, especially from complex databases, can be very time consuming, often taking weeks to months, while occupying up to 80% of the overall effort in AI and machine learning projects.

Join Thanh Lam Hoang and Beat Buesser from the IBM Research team as they discuss a new capability in Watson Machine Learning for z/OS to automate feature engineering for databases with multiple tables. They will provide a step-by-step demonstration and share examples of machine learning models where accuracy improved significantly after applying this new technique. This automated approach to feature engineering can help generate machine learning models from raw data, competing head to head with even the most experienced data scientists.

Automated feature engineering can make both novice and expert data scientists more productive. It can help reduce overall efforts in machine learning initiatives and enhance the  model development life-cycle.  Register today to join the presentation!