Astronomy 598 Topics in Theoretical Astrophysics (Machine Learning for Computational Science)

Winter 2017: Fri 11:00-12:20 Physics/Astronomy Building A210

Instructor: Pramod Gupta
psgupta *at* astro. washington. edu

Office hours: after class, or email.


Students should know a programming language. Students should bring their laptop to class. If you have a windows laptop please install ssh client software on your laptop so that you can connect to hyak. Students should have a hyak account and should bring the e-token to class. Any UW student can get a hyak account. Instructions are at below link:


(1) Linear Models for regression (Maximum Likelihood Estimation, Least-squares)

(2) Linear Models for classification (Linear Discriminant Analysis, Logistic Regression, Maximum Likelihood Estimation, Gradient Descent)

(3) Linear Models and Regularization (Ridge Regression, Lasso)

(4) Kernel density estimation and classification (Naive Bayes Classifier)

(5) K-nearest neighbors and classification

(6) Model Complexity (Overfitting, underfitting, Bias-Variance tradeoff)

(7) Resampling (K-fold Cross-validation, Bootstrap)

(8) Cluster Analysis (hierarchical clustering, K-means, Gaussian mixtures)

(9) Principle Component Analysis

(10) Singular Value Decomposition


Homework problems (100% of grade) will be assigned every week.


The Elements of Statistical Learning
Second Edition
Hastie, Tibshirani, Friedman