HST951 - Medical Decision Support - Spring 2001
Date Lecture Title Lecture content Reading (author) Homework Lecturer
2/6 Overview of Course

ppt file
  • Handouts
   Ohno-Machado
2/8 Decision Analysis I
  • Bayes rule
  • Decision Trees
  • Value of information
  • Utilities
 Col
2/13 Decision Analysis II
  • Sensitivity Analysis
  • Intro to Markov Models
  HW#1: Exercises 1 to 5 in Shortliffe's
Chapter 3 due on 2/20		 Col
2/15 Probability and Decition Theory
  • Conditional Independence
 Szolovits
2/22 Bayesian Networks I

Video
  •  Principles
  • Graphical Representation
   Szolovits
2/27 Bayesian Networks II Ramoni
3/1 Learning from longitudinal data   Homework 2 (due on 03/13) Ramoni
3/6 Logistic Regression I
  • Contingency tables
  • Logistic function
  •  PROC LOGISTIC in SAS
  • Class notes
  • NEMJ on DHCA vs. Low-flow bypass (Newburguer)
 Wypij
3/8 Logistic Regression II
  • Odds-ratios
  • Variable selection
  •  APACHE paper (Knaus)
  • NN + GA (Dybowski)
   Wypij
3/13 Classification Trees & CART
  • Analogy to logistic regression
  • Overfitting
  • Applications
  • Intro Class.Tree (Winston's Chapter)
  • Trees for FNAB (Cross)
 Homework 3
(due 3/20)		 Long
3/15

 

 Rough Sets
  • Entropy
  • Gain
  • Pruning
  • Comparing C4.5 and CART
   Vinterbo
3/20

 

 Neural networks

ppt file
  • Boolean logic
  • Equivalence classes
  • Intro to NN (Ohno-Machado)
  • Diagnosis of MI (Baxt)
 Ohno-Machado
3/22

 

 Evaluation Methods I

ppt file

 Homework 4  Ohno-Machado
4/3

 

 Case Selection and Sampling

ppt file

 Ohno-Machado
4/5 Variable Selection   Vinterbo
4/10 Application at BWH Bates
4/12 Clustering Techniques

ppt file

   Ohno-Machado
4/19 Internist  Szolovits
4/24 Bagging and Boosting

ppt file

   Ohno-Machado
4/26 Anonymization and Privacy
  • Definitions
  • Current systems
  • Blocking Inference
   HW#9: NNeighbor and Clustering Ohno-Machado
Vinterbo
5/1 Unsupervised Learning
  • Clustering
  • Metrics
  • Exploratory data analysis methods
  • SOM
  • R.Shepard's paper
  • Kohonen's paper
   Kim
5/3 Clinical Alerts and Reminders   Gill Kuperman
5/8 Cluster Quality Measures   Kim
5/10 Designing Clinical Trials    Laura Smeaton 
5/15 Student presentations  15 minute presentation 
(plus 5 minutes for questions) 
Include: (1) Introduction/Motivation/Background, (2) Material and Methods, (3) Results, (4) Conclusion, (5) Discussion		    Do not forget to report results using evaluation methods taught in class (ROC, calibration curves, etc.)...   Ohno-Machado
5/17 Student presentations    FINAL REPORTS DUE  Ohno-Machado

Textbooks for reference (all optional):


Last updated 1/10/01