Lecture Materials

Lecture 1: Introduction to Machine Learning (Jan. 12 ) (slides)

• • optional reading

  • textbook section: ISL chapter 2

  • Applications of Machine Learning

    • • Autonomous Driving (video)

      • "How Credit Card Companies Spot Fraud Before You Do" (article, U.S. News)

      • "Mining Electronic Records for Revealing Health Data" (article, New York Times)

      • "The Algorithm That's Hunting Ebola" (article, IEEE Spectrum)

  • optional R tutorial (video, 25 minutes)

  • tutorial code

  • tutorial datasets: Auto.data, Auto.csv

Lecture 2: Unsupervised Learning, Clustering and Dimensionality Reduction (Jan. 14) (slides)

• • optional reading

  • ISL: Chapter 10 (PCA 10.2, clustering 10.3)

    • ISL: Section 10.4 "Lab 1: Principal Component Analysis"

  • optional review of eigenvectors and eigenvalues (video, 10 minutes) (slides)

Lecture 3: Dimensionality Reduction Part II (SOM, MDS, ICA) and Imputation (Jan. 19) (slides)

• • optional reading

  • ESL: 14.4 (Self-Organizing Maps)

  • ESL: 14.8 (Multidimensional Scaling)

    • "America’s Broken Politics” (article, New York Times)

    • http://voteview.com/polarized_america.htm

  • ESL: Section 14.7.2 (ICA, advanced topic, skim only)

    • • paper on ICA (optional, more advanced treatment, skim sections 1 "Motivation" and 7 "Applications of ICA")

  • ESL: Section 9.6 (imputation)

• optional R demonstration of SOM and PCA code examples (video, 15 minutes)

  • • PCA code in R and corresponding dataset (chemical levels in olive oils from different regions)

    • SOM code in R (OS X users must first install X Quartz)

Lecture 4: Unsupervised Learning Wrap-up (NMF) & Supervised Learning Intro (Jan. 21) (slides)

• • optional reading

  • ESL: Section 9.6 (imputation)

  • ESL: Section 14.6 (NMF)

    • ISL: Chapter 3 (linear regression)

  • ISL: Section 4.3 (logistic regression)

Lecture 5: Cross-validation, Regularization (lasso, ridge, elastic net, PCR), and Sparsity (Jan. 26) (slides)

• • optional reading

  • ISL: Section 5.1 (cross-validation)

    • • "Many Psychology Findings Not as Strong as Claimed, Study Says" (article, New York Times)

• • • ISL: Section 6.2 (lasso and ridge regression)

Lecture 6: Support Vector Machines (SVM) (Jan. 28) (slides)

• • optional reading

    • ISL: Chapter 9 (support vector machines)

    • Andrew Ng's CS 229 course notes on SVMs available here for those interested in the mathematics behind SVMs.

Lecture 7: Classification and Regression Trees (CART) and the Bootstrap (Feb. 2) (slides)

• • optional reading

    • ISL: Section 8.1, 8.3.1, and 8.3.2 (CART using "tree")

    • ISL: Section 5.2 and 5.3.4 (The Bootstrap)

  • optional R demonstration of CART tree: "rpart" command, an alternative to "tree" (video, 10 minutes)

  • R Code for CART example and corresponding dataset

Lecture 8: Ensemble Methods (Boosting, Bagging, Random Forest) and Neural Nets (Feb. 4) (slides)

• • optional pre-lecture reading

    • ISL: Section 8.2.1, 8.2.3, and 8.3.4 (Boosting and Bagging)

    • ISL: Section 8.2.2 and 8.3.3 (Random Forests)

    • ESL : Chapter 11 (Neural Networks)

  • Interpretability of Deep Neural Networks:

    • Google's "Deep Dream"

    • Cats on YouTube (research paper) (NYTimes)

    • Interview with a Top Kaggle Competitor

ISL refers to the course textbook An Introduction to Statistical Learning with Applications in R

ESL refers to the course reference book The Elements of Statistical Learning