Business Analytics: Decision Making Under Uncertainty

This set of vocabulary flashcards covers the fundamental concepts of Decision Analysis, including the components of decision trees, EMV calculations, multistage problems, and Bayesian probability updating as presented in the Business Analytics lecture notes.

Decision Analysis Framework

A formal framework for analyzing decision problems involving uncertainty, including criteria for choosing alternatives, use of probabilities, the effect of early decisions on later ones, quantifying the value of information, and the effect of risk attitudes.

Decision Tree

A powerful graphical tool that enables a decision maker to view decision alternatives, uncertain outcomes and their probabilities, economic consequences, and chronological order of events at once.

Decision Node

Represented by a square in a decision tree, it indicates a point in time when the decision maker makes a choice.

Probability Node

Represented by a circle in a decision tree, it represents a time when the result of an uncertain outcome becomes known.

End Node

Represented by a triangle, it indicates that the problem is completed, all decisions have been made, all uncertainty has been resolved, and all payoffs and costs have been incurred.

Expected Monetary Value (EMV)

The weighted average of the possible payoffs for a decision, weighted by the probabilities of the outcomes.

EMV Criterion

A decision-making approach where the option with the largest EMV is chosen, sometimes called "playing the averages."

Folding-back Process

A procedure for finding EMVs and the optimal decision by starting from the right of the tree and working back to the left, calculating EMVs at chance nodes and selecting the maximum EMV at decision nodes.

One-stage Decision Problem

A problem where only one decision is made at a single point in time.

Multistage Decision Problem

A decision problem that evolves through time in stages, where early decisions and uncertain outcomes affect later decisions and probabilities.

Contingency Plan

An EMV-maximizing strategy for multistage problems that specifies which decision to make at each stage based on previous uncertain outcomes.

PrecisionTree

A spreadsheet add-in by Palisade Corporation that allows users to draw decision trees, perform folding-back automatically, and conduct sensitivity analysis.

Sensitivity Analysis

A process used to see whether the best decision changes as one or more input parameters change.

Risk Profile

The full probability distribution of monetary outcomes for a chosen strategy, available through PrecisionTree analysis.

Bayes' Rule

A formal mathematical mechanism for updating probabilities as new information becomes available.

Prior Probabilities

The original probabilities assigned to outcomes before new information is observed.

Posterior Probabilities

The updated set of probabilities for outcomes after new information is observed, denoted as $P(O|I)$, which must sum to $1$.

Likelihoods

The conditional probabilities of observing specific information given a certain outcome, denoted as $P(I|O_n)$, which are used as inputs for Bayes' rule.

Law of Total Probability

The denominator in the Bayes' rule formula, representing the total probability of the information outcome $P(I)$.

Strategy Region Graph

A chart showing how the EMV varies with conditions to help identify if the optimal decision changes over the range of an input variable.

Two-way Sensitivity Chart

A chart showing how the selected EMV varies as each pair of inputs varies simultaneously.