- This event has passed.
Friday August 17 @ 8:30 am - 12:00 pm
Supervised Machine Learning: Still the Most Important Machine Learning Technique
with Keith McCormick, Senior Consultant and Trainer @ The Modeling Agency
Regression, decision trees, neural networks – along with many other supervised learning techniques – provide powerful predictive insights. New users of these established techniques are often impressed with how easy it all seems to be since automated model-building software is widely available. However, proper data preparation and human input is still necessary to get optimal results.
We will dedicate about half of the time on practicing translating the business problem into a form the algorithms can assist with and about half learning why preparing the data for optimal performance during modeling is still largely a human activity. If your organization is fairly new to predictive analytics it is unlikely that any other technique deserves more attention than getting well versed in the basics of Supervised Machine Learning.
You Will Learn
- When to apply supervised modeling methods
- Options for inserting machine learning into the decision making of your organization
- How to properly prepare data for different kinds of supervised models
- Explore the pros and cons of “black box” models including ensembles
From Unsupervised to Supervised Learning
with Srihari Nagarajan, Senior Data Scientist @ Metlife
Companies use anomaly detection techniques to identify and mitigate fraud. Effective use of machine learning techniques could help save millions of dollars. At the same time, there is a cost involved in having investigators analyze each and every case to hand-classify them as fraud or not. Semi-supervised learning could pave way to reduce this cost greatly by minimizing human efforts. Once we believe that we have captured most of the fraud patterns in the data, we could use supervised learning techniques to easily identify fraud.
You will learn
When to apply anomaly detection techniques
How to transition to semi-supervised/supervised learning