CAP Bayesian Networks and Decision Flashcards
Bayes’ Theorem
P(A|B) = P(B)P(B|A)/P(A); posterior equals prior times likelihood over marginal
Bayesian Network
a probabilistic graphical model that represents a set of random variables and their conditional dependencies via a directed acyclic graph
Decision analysis
a quantitative method which permits the systematic evaluation of the costs and/or benefits accruing to courses of action that might be taken in a complex problem
Decision matrix
a list of values in rows and columns that allows an analyst to systematically identify, analyze, and rate the performance of relationships between sets of values and information
Decision making under certainty
Using the state of nature to determine which outcome in the decision matrix is best for that state of nature, then deciding in order to achieve that outcome
Maximax return
Optimistic decision making; choose the best possible outcome for each decision, then choose the best result from that group
Maximin return
Pessimistic decision making; choose the worst possible outcome for each decision, then choose the best result from that group
Minimax Regret
Decision making based on potential loss from best outcome; subtract each outcome from the best outcome possible for that decision, then choose the result with the smallest different/regret
Decision tree
A way to structure multi-stage decision problems that allows us to identify courses of action, consider probability of multiple different states of nature, assess outcomes, and select the best alternative
Expected value of decision tree
Method for using decision trees that calculates the sum of the products of the probabilities and outcome values
Value focused thinking
Method for making decisions that focuses first on articulating values, then applying values to quantification of objectives
Pareto efficient frontier
The set of all resource allocations such that it is impossible to make one preference criterion better without making at least one preference criterion worse
