Probability and Statistical Data Analysis: Hypothesis Testing and Machine Learning Model Validation
Master the probabilistic reasoning and statistical inference that machine learning is built on. Part five of a five-part series, this online course teaches you how to test hypotheses and choose the right machine learning models for deployment.
Course Information
Certificate Track
Learn for Free
About this Course
Test scientific hypotheses with regression, validate assumptions with diagnostics, and control false discoveries across multiple comparisons. Then practice the modeling workflow with validation, interpretability, calibration, and deferral strategies. By the end of this course, you will have the knowledge and experience with an end-to-end workflow for building, evaluating, and deploying models you can defend with confidence.
This course is part five of five in the Probability and Statistical Data Analysis Program, which includes the following:
- Foundations and Exploratory Data Analysis
- Discrete Distributions and Categorical Data
- Multivariate Statistics and Bayesian Inference
- Uncertainty Quantification and Data-Driven Decisions
- Hypothesis Testing and Machine Learning Model Validation
What you'll learn
- Speak the language of sample spaces and events
- Reason conditionally about how outcomes relate
- Turn a scientific question into a probability question
- Summarize data with histograms, boxplots, and proportions
Meet your instructors
Devavrat Shah
Andrew (1956) and Erna Viterbi Professor, Electrical Engineering and Computer Science; Director, MicroMasters Program in Statistics and Data Science
Devavrat Shah is the Andrew (1956) and Erna Viterbi Professor of Electrical Engineering and Computer Science, a member of IDSS and LIDS at MIT, and an IEEE Fellow. He received his PhD in computer science from Stanford University in 2004 and joined the MIT faculty in 2005. His research spans statistical inference, stochastic networks, and social data processing, driven by the challenge of designing simple, high-performance algorithms under severe resource constraints. Dr. Shah is co-founder of the machine learning startup Ikigai Labs.
