TIM209, Winter 2014, Section 01: Lectures

Slides contain lecture slides PLUS homework.

TIM 209 Winter 2014, Lecture Schedule

Date Slides+HW (HW at end of slides) SupplementaryMaterial Week Lecture Topics
1/9/14

Slides+HW Lecture 1

 
1 Introduction to class, Intro to data Science, Data mining, Machine Learning, Statistical from summaries to decision-making systems, digital Advertising 101, Obama's ’Data Science' Victory, recommender systems al Netflix, lifelogging, R
1/16/14  Slides + HW for Lecture 2

 Ref Material for class and class notes  

 

2 6 steps in data modellig; simple linear regression, multiple linear regression, linear regression in R, lienar regression diagnostis, worked examples of linear regression, brute force linear regression, closed form linear regression, error surfaces 
1/23/14 Slides + HW for week 3
  3 Objective function, Gradient function, second derivatives, zeros of a gradient function, bisection algorithms, Newton-Rhapson algorithm, Gradient descent, linear regression via gradient descent (an approximation of Newton-Rhapson algorithm)
1/30/14 Slides + HW for week 4   4 Classification versus Regression, Probability Theory Review Naïve Bayes (Generative Models),  Feature Selection, Metrics
2/6/14 See Dropbox Folder   5 Bernouilli Naive Bayes, Multinomial Naive Bayes, Regularization, LASSO, Ridge regression
2/13/14 See dropbox folder (MidTerm-Week6)    6

Review of Bernouili Naive Bayes and other questions.

Midterm exam is on Sunday 2/16/2014 from 6PM to 9PM. Please see class dropbox folder: MidTerm-Week6 for details.

2/20/14 See Dropbox Folder for slides, Homework and referecne material   7

Feature Selection, Metrics, Logistic Regression (LASSO, Ridge Regression again!), Kernel Density Estimation Decision Trees

2/27/14 See Dropbox Folder for slides, Homework and reference material   8

Regression Trees, Classification Trees, Ensembles, Bagging, Random forests

3/6/14 See Dropbox Folder for slides, Homework and reference material   9

Data Science Competitions, Heritage healthcare prize (HHP), Feature engineering for HHP, preparing a submission for HHP, code in R