Practical Predictive Analytics: Models and Methods

Data science is all about statistical experiment design and analysis. This course will teach you how to design and analyze statistical experiments using modern methods. The course will cover common pitfalls when interpreting statistical arguments, particularly those involving big data. This course, taken together, will allow you to internalize a core set practical and efficient machine learning concepts and methods, and then apply them to real-world problems.

Learning Objectives: You will be able: You will be able to design and analyze experiments that produce effective results. To make bulletproof statistical arguments, you can use resampling techniques without resorting to esoteric notation. Explain and apply a core collection of classification methods increasing in complexity (rules trees random forests) and the associated optimization methods (gradient descend and variants). Explain and apply unsupervised learning concepts and methods. Explain the most common terms in large-scale graph analytics. This includes structural query, traversals, recursive queries and PageRank.