@ Online event
This talk will cover two classic approaches to a common problem in data analysis: inferring unobserved groups, or clusters, in data. This is a common task, for example in discovering subpopulations in data of people (e.g. customer segmentation), compressing data, or finding low-dimensional representations of data. In particular we will examine K-means clustering with Lloyd's algorithm and Gaussian mixture modeling with the expectation-maximization algorithm. Implementation with visualization will be provided.
About the Speaker:
William Grisaitis (@grisaitis) is a machine learning engineer and data analyst. Originally from Orlando, he has worked in a variety of roles in analysis and engineering at Capital One, in a ML research lab, and as a consultant for startups and small companies. William graduated with a B.A. from Duke University in 2012.