Mathematical Methods in Data Science

Sébastien Roch is a Vilas Distinguished Achievement Professor of Mathematics at the University of Wisconsin, Madison. At UW-Madison, he helped establish the Data Science Major and has developed several courses on the mathematics of data. He is the author of Modern Discrete Probability: An Essential Toolkit (2023).

• 1. Introduction: a first data science problem; 2. Least squares: geometric, algebraic, and numerical aspects; 3. Optimization theory and algorithms; 4. Singular value decomposition; 5. Spectral graph theory; 6. Probabilistic models: from simple to complex; 7. Random walks on graphs and Markov chains; 8. Neural networks, backpropagation and stochastic gradient descent.

'Mathematical Methods in Data Science provides a clear and accessible primer on key concepts central to data science and machine learning. Through engaging examples from neural networks, recommender systems, and data visualization, Roch illuminates myriad foundational topics and methods. Designed for readers from a broad range of backgrounds, this text is an indispensable resource for students and professionals.' Rebecca Willett, University of Chicago