In this blog you will learn that the less probable the interesting event is, the more data it takes to obtain enough to generalize a model to unseen cases, and why some projects probably should not proceed until enough critical data is gathered to make them worthwhile.
BLOG: Fraud detection is about finding needles in haystacks and requires reliably labeled instances of fraud and non-fraud behavior to train a predictive model to best separate fraud from non-fraud cases. But what do we do when labels are not just rare, but are completely absent?
This blog explores the difficulty knowing how to get started using data science and predictive analytics and how choosing the right problem and focusing on a few key guidelines delivers greater business value and gains support for analytics from key stakeholders.
Elder Research Data Scientist Stuart Price discusses Goodhart’s Law and the risk that any metric applied to a competitive or adversarial system will change behavior. If your adversary has a good chance of figuring out your metric, how can you keep your system from being gamed?
Blog discusses a 6-step communication framework to build team consensus and overcome resistance to the disruptive changes introduced by advanced analytics.