PSYCH 4508 Lecture Notes - Lecture 3: Sunk Costs, Decision Tree Learning, Decision Aids
Document Summary
Lecture: august 31, september 2 & september 7. Picturing decisions with decision trees: a decision tree is a graphical model of a particular decision situation. Describes the decision maker"s relevant beliefs and values: normative model. What would be the rational decision: prescriptive model. A technique or tool for thinking systematically about the decision: not usually descriptive model. Most decisions are not made by using decision trees. Identify all of the decision alternatives (choice options) For each alternative, what are the uncertain events and the possible consequences: beliefs. How likely (probable) are each of the consequences: utilities (values) Choose numbers for the best and worst consequences. Relative to those endpoints, how good are the other consequences: expected utilities. Choose the alternative with the highest eu (if following eu rigorously: decision trees also work with expected values (evs) At each chance node, the probabilities should sum to 1. 0.