Propensity Measurement Using Positive-Unlabelled Bagging

Explore propensity measurement techniques using Positive-Unlabelled (PU) bagging in this 26-minute conference talk from the Data Science Festival. Dive into the semi-supervised binary classification method that trains on datasets where only positive instances are labeled, while the remaining instances are unlabeled or unknown. Learn from speaker Samir Bajaj as he discusses the intricacies of PU learning and its applications in propensity measurement. Gain insights into this powerful technique that can be particularly useful when dealing with partially labeled datasets in various data science and machine learning scenarios.