Decision Field Theory:
A Stochastic-Dynamic Model of
Human Deliberation
IEE 511 Analysis of Decision Processes
2011, Ronald G. Askin
Decision Field Theory
Rational theorists (economists) goal: logical models for ideal DMs preferences.
Behavioral scientis
Signal Detection Theory
IEE 511 Analysis of Decision Processes
2011, Ronald G. Askin
Signal Detection Theory (SDT)
SDT is a model for making decisions for discriminating between different
classes of items/states of nature under uncertainty.
Used in Eng
Behavioral Decision Making:
What Real People Do
IEE 511 Analysis of Decision Processes
2011, Ronald G. Askin
Behavioral Decision Making
Support for rationality
1. Try to maximize the chance of achieving goals when making
choices.
2. Competition should fa
VIII. Multiattribute Decision
Making Under Uncertainty
IEE 511 Analysis of Decision Processes
2011, Ronald G. Askin
A. General Properties and
Background Results
IEE 511 Analysis of Decision Processes
2011, Ronald G. Askin
2
Note: Need Only Consider Pare
Common Human Biases in
Decision Making
IEE 511 Anal. of Decision Processes
Human Bias in Decision Making
Common Biases
Anchoring
Certainty Effect
Cognitive Dissonance
Confirmation
Congruence
Dominance or Focusing
Framing
Herding
Hindsight
Hypothesis/Expec
V. Utility Theory:
Prescriptive Behavior for the Rational
Economic Human
IEE 511 Analysis of Decision Processes
2011, Ronald G. Askin
Utility Theory
Loss functions indicated the admissibility of decision rules that
accommodate nonlinear perceptions at d
The Bayesian Way: Prior Beliefs Count
1
IV. C. Bayesian Decision Models
A. Bayesian Updating
In general, we have a random process for variable X with pdf
p(X = x ) which has some distributional form for X with
parameters . is unknown so we adopt a prior
New Product Development
Problem:
Pursue R&D or maintain same product portfolio or do minor updates to
current products
If R&D succeeds, should we market or license
Uncertainties
Will R&D succeed
If R&D succeeds technologically, will new product pene
Key Distributions and
Subjective Probability Assessment
IEE 511 Analysis of Decision Processes
2011, Ronald G. Askin
Review of Probability
The Binomial Distribution
The probability of r successes in n independent
trials given the probability of success
III Graphical Tools for Complex Scenarios
A. Influence Diagrams (Relevance Diagram)
B. Decision Trees
IEE 511 Analysis of Decision Processes
2011, Ronald G. Askin
A. Influence Diagrams
Also called Relevance Diagram
A snapshot of the Decision Makers pre
Analytical Hierarchy Process (AHP)
Founded by T. L. Saaty in 1977, J. of Mathematical Psychology; 1980
The Analytical Hierarchy Process.
A popular and widely used method for multi-criteria decision making
under Certainty.
Allows the use of qualitative, as
Multiattribute Comparison of
Alternatives Under Certainty
(Tradeoff Studies)
IEE 511 Analysis of Decision Processes
2011, Ronald G. Askin
Choosing Between Multiple
Alternatives/Options with Multiple Criteria
M criteria, i = 1, M
N Alternatives, j = I, ,
Multiattribute Value Functions
(Pseudo Utility Theory for Deterministic Outcomes)
Definitions
Each alternative j=1,n has a measured value xij
criterion i = 1,M
for each
Multiattribute Value Function V ( x1 , x2 ,., xM )
Additive Value Function V ( x1 ,
Review of Probability
(and a little more)
IEE 511 Anal. of Decision Processes
Ronald G. Askin
IEE 511 Anal. of Decision Processes
Ronald G. Askin
2
Basic Definitions
Random Experiment: an experiment in which the outcome can not be
known in advance with ce
IEE 511 Analysis of Decision
Processes
Ronald G. Askin, Professor and Director
School of Computing, Informatics and Decision
Systems Engineering
Arizona State University
[email protected]
IEE 511 Anal. of Decision Processes
Ronald G. Askin, 2009
When you
Chapter 14. Principal Component
Analysis
Nong Ye
Data Mining: Theories, Algorithms, and Examples
1
Overview
Principal component analysis is usually used to
reduce the dimensionality of data so that the
data can be further visualized or analyzed in a
low-
Chapter 20. Wavelet Analysis
Nong Ye
Data Mining: Theories, Algorithms, and Examples
1
Overview
Many objects have a periodic behavior and thus show a
unique characteristic in the frequency domain
Human sounds different from some animal sounds in
frequen
Chapter 12. Association Rules
Nong Ye
Data Mining: Theories, Algorithms, and Examples
1
Overview
Association rules uncover items that are
frequently associated together
Market basket analysis uncover what items
customers often purchase together
Nong Ye
Chapter 13. Bayesian Network
Nong Ye
Data Mining: Theories, Algorithms, and Examples
1
Overview
A Bayes classifier requires all the attribute
variables are independent of each other
A Bayesian network allows associations among
the attribute variables th
Chapter 5. Artificial Neural Networks
for Classification and Prediction
Nong Ye
Data Mining: Theories, Algorithms, and Examples
1
Overview
Artificial Neural Networks (ANNs) use the
basic architecture of human brain which
consists of neurons and connectio
Chapter 9. K-Means Clustering
and Density-based Clustering
Nong Ye
Data Mining: Theories, Algorithms, and Examples
1
Overview
K-means and density-based clustering
algorithms produce non-hierarchical groups of
similar data points based on the centroid and
Chapter 8. Hierarchical Clustering
Nong Ye
Data Mining: Theories, Algorithms, and Examples
1
Overview
Hierarchical clustering produces groups of
similar data points at different levels of
similarity
Nong Ye
Data Mining: Theories, Algorithms, and Examples
Chapter 4. Decision and
Regression Trees
Nong Ye
Data Mining: Theories, Algorithms, and Examples
1
Overview
Decision and regression trees learn and
express classification and prediction patterns
in the form of a tree
A decision tree classifies the categ
Solving Recurrences
Use a recurrence to characterize the running time of a divide-andconquer algorithm, and solving the recurrence gives us the
asymptotic running time
Examples of recurrence
Chapter 4. Divide-and-Conquer
IEE 605, Nong Ye
4.3 The Substitut
Chapter 1. Introduction to Data,
Data Patterns and Data Mining
Nong Ye
Data Mining: Theories, Algorithms, and Examples
1
Overview
Data mining aims at discovering useful data
patterns from massive amounts of data
1.1 Examples of Small Data Sets
1.2 Type
Insertion Sort: Algorithm Design
Insertion sort algorithm
An incremental approach is used in INSERTION-SORT: having
sorted A[1j-1], place A[j] correctly, so that A[1j] is sorted
Chapter 2. Getting Started
IEE 605, Nong Ye
Insertion Sort: Algorithm Design
Asymptotic Notations
Asymptotic efficiency: focus on the order of growth when input sizes
are large
An asymptotic-efficient algorithm is usually the best choice for all but
very small inputs
Chapter 3. Notations
IEE 605, Nong Ye
Asymptotic Notations
O-not
Algorithms
Algorithm: a sequence of computational steps that transform the input
into the output or solve a well-specified computational problem
An example of a computational problem: a sorting problem:
Input: a sequence of n numbers a , a , , a
1
2
n
O