About Stat 154

Bin Yu

What is Stat 154?

This course aims at training students to solve real world prediction problems. We achieve our goal through data experience and training of critical thinking, use of domain knowledge, data visualization, machine learning algorithms and mathematics. We take a holistic view of prediction in the data science life cycle that consists of problem formulation, data collection, exploratory data analysis (EDA), unsupervised learning, supervised learning, data results, validation, and conclusion.

We separate the real world from the world of mathematics and algorithm, and cover systematic methods for seeking evidence to connect the two worlds (sometimes well but often not). In particular, to help connect the two worlds, we emphasize the concepts PQR-S: P for population, Q for question, R for representativeness and S for scrutiny, and we follow the PCS framework: P for predictability, C for computability and S for Stability.

Stat 154 prerequisites

Mathematics 53 or equivalent; Mathematics 54, Electrical Engineering 16A, Statistics 89A, Mathematics 110 or equivalent linear algebra; Statistics 133, 134; Statistics 135 or equivalent; experience with some programming language. Recommended prerequisite: Mathematics 55 or equivalent exposure to counting arguments.

What is the best way to succeed in Stat 154?

Stat 154 is an upper division course that requires both mathematical and programming backgrounds. Students are expected to have a corresponding degree of independence, discipline, and mathematical maturity.

Will Stat 154 be offered in both Fall 2019?

Yes.