Analytics I: Statistical Learning
A first course on the theory and applications of statistical signal processing. The material benefits students interested in the design and analysis of signal processing systems, with numerous examples illustrating theory and applications such as high-resolution spectral analysis, system identification, digital filter design, adaptive beamforming and noise cancellation, and tracking and localization.
This course dives into the mathematics behind data analysis and machine learning, unveiling cutting-edge tools in statistics and signal processing. Learn essential regularization techniques for predictive modeling, applied across finance, marketing, social networks, and engineering. With PyTorch experiments, you will grasp machine learning concepts hands-on.
Prerequisites: Background in probability and random processes, linear and matrix algebra, and exposure to basic signal processing.
Looking for the current offering? See Analytics I: Statistical Learning (2026).
Course materials
- Syllabus
- 0. Introduction
- 1. Probability and stationary processes
- 2. Eigen analysis, SVD, PCA and matrix completion
- 3. The learning problem
- 4. Training vs testing
- 5a. The linear model and optimization
- 5b. Nonlinear transformation and logistic regression
- 6. Overfitting and regularization
- 7. Lasso regression
- 8. Support vector machines
- 9. Neural networks
- 10. Convolutional neural networks
- 11. Deep autoencoders and generative models
- 12. Transformer models
- 13. Diffusion models
- 14. Graph neural networks
- 15. Epilogue
- Supplementary material
- Video