Unformatted text preview: nd @RISK
Excellent software exists to help a
decision analysis fit theoretical
distributions to data. BestFit, a program published by
Palisade Software. Providing a way for
an analyst to fit theoretical distributions
to subjective assessments elicited from
experts. 18 CHAPTER 10. USING DATA Slide No. 36
ENCE 627 ©Assakkaf Software for Fitting Distributions:
BestFit and @RISK
Fitting distributions to the halfway
house data with BestFit. Read procedure of BestFit with @RISK
packages in book pages 405 – 411. CHAPTER 10. USING DATA Summary Slide No. 37
ENCE 627 ©Assakkaf We have seen some ways in which data
can be used in the development of
probabilities and probability distributions
for decision analysis.
The basis of constructing histograms
and empirically based CDFs.
Use of data to estimate parameters for
theoretical distributions. 19 CHAPTER 10. USING DATA References Slide No. 38
ENCE 627 ©Assakkaf 1. Berry, D. (1995). Statistics: A Bayesian Perspective,
Belmont, CA: Duxbury. 2. Chatterjee, S., and B. Price (1977) Regression Analysis by
Example. New York: Wiley. 3. DeGroot, M. (1970) Optimal Statistical Decisions. New
York: McGrawHill. 4. Draper, N., and H. Smith (1981) Applied Regression
Analysis, 2nd Ed. New York: Wiley. 5. Neter, T., W. Wasserman, and M. Kutner (1989). Applied
Linear Regression Models, 2nd Ed. Homewood., IL: Irwin. 6. Olkin,I., L.J. Gleser, and C. Derman (1980) Probability
Models and Applications. New York: Macmillan. 7. Raiffa, H., and R. Schlaifer (1961) Applied Statistical
Decision Theory. Cambridge, MA: Harvard University CHAPTER 10. USING DATA Slide No. 39
ENCE 627 ©Assakkaf Using Data to Model Relationships Using Data
to
Model Relationships 20 CHAPTER 10. USING DATA Slide No. 40
ENCE 627 ©Assakkaf Using Data to Model Relationships
We use data to try to understand the
relationships that exist among different
phenomena in our world.
Examples:
– Causes of Cancer
– Sales Revenue
– Economic Conditions
– Natural Processes CHAPTER 10. USING DATA Slide No. 41
ENCE 627 ©Assakkaf Using Data to Model Relationships
Note:
1. The motivation for studying relationships among
phenomena that we observe is to gain some
degree of control over our world. In many cases
we hope to make changes in those areas where
we have direct control in order to accomplish a
change in another area. 2. We will focus on the problem of using data on a
number of auxiliary variables (which we will
denote as X1, …, Xk) to determine the distribution
of some other variable of interest (Y) that is
related to the x’s. 21 CHAPTER 10. USING DATA Slide No. 42
ENCE 627 ©Assakkaf Using Data to Model Relationships 3. Y is sometimes called a response
variable or dependent variable, because
its probability distribution changes in
response to changes in the X’s.
4. The X’s sometimes are called
explanatory variables or independent
variables, because they can be used to
help explain the changes in Y. CHAPTER 10. USING DATA Slide No. 43
ENCE 627 ©Assakkaf Using Data to Model Relationships – The use of data to understand relationships is
not trivial. Consider the influence diagrams.
• The brute force approach would require obtaining
enough data to estimate the conditional distribution
for the particular variable of interest (Y...
View
Full Document
 Fall '12
 MansoorehMollaghasemi
 Regression Analysis, Probability theory, Yi, SLIDE NO.

Click to edit the document details