Rada Menczel - The Secret Sauce of Data Preprocessing in Machine Learning

Abstract:

When data scientists want to train new models, they have a general idea of how their flow will look like. Assuming that the problem they need to solve is well defined, they need to explore the data, define labels, visualize, train, evaluate, tune and test. The most time consuming and often tedious part is data preprocessing and preparation. If you don’t invest enough in this, you may still get a decent or even a good model, but is that enough? What if I told you that there is a case that is often unnoticed by data scientists? And by adding this small step you can improve your model results and get a great model?  In this talk I’ll present a problem that is often being ignored - identical feature vectors with different labels, why this happens, and how you can solve it in different ways in all possible domains. After this talk you will add this phase to your preprocessing toolkit and won’t look back.

Bio:

Rada Menczel, Head of Machine Learning

Rada Menczel is data science manager. She has vast experience as head of machine learning and data science director in cyber security, fin-tech and e-commerce companies, where she led teams of data scientists and was responsible for all aspects of the ML models in the company, from POC and researching algorithms and models to production.

Rada has an MSc in Information Systems Engineering from BGU with specializations in machine learning and recommender systems. She is enthusiastic about data science, machine learning, deep learning and basically anything that is related to learning.

More about the speaker: https://www.womenonstage.net/speakers/Rada-Menczel