|
Applied
Regression Analysis for Business Decision Making
Seminar
Description: This
course aims to familiarize the student with statistical
prediction. It covers simple and multiple regression methods as
well as time series and forecasting models in business. Instead of
theoretical development, the course emphasizes the application of
these methods in business systems analysis and improvement.
Prerequisites:
Participants
in this seminar must have achieved Six Sigma Yellow Belt
certification.
Six
Sigma Certification:
Applied Regression Analysis for
Business Decision Making is required to sit for the Six Sigma
Black Belt examination.
Textbook:
Montgomery, D.C., Peck, E. A., Vining, G. G., Introduction to
Linear Regression Analysis, 3rd Edition, Wiley and
Sons.
Statistical
Software:
Minitab 14
Seminar
Outline:
Chapter
1: Introduction
Regression and
Model Building
Data Collection
Uses of Regression
Chapter
2: Simple Linear Regression:
Simple Linear
Regression Model
Least Squares
Estimation of Parameters
Hypothesis testing
on the Slope and Intercept
Interval Estimation
in Simple Linear Regression
Prediction of New
Observations
Coefficient of
Determination
Considerations
Regression Through
the Origin
Estimation by
Maximum Likelihood
Cases Were the
Regressor is Random
Chapter 3: Multiple
Linear Regression
Multiple Regression
Models
Estimation of Model
parameters
Hypothesis testing
in Multiple Linear Regression
Confidence in
Multiple Regression
Prediction of New
Observations
Hidden
Extrapolation in Multiple Regression
Standardized
Regression Coefficients
Multicollinearity
Why Do Regression
Coefficients Have the Wrong Sign?
Chapter
4: Model Adequacy Checking
Residual Analysis
PRESS Statistic
Detection and
treatment of Outliers
Lack of Fit of the
Regression Model
Chapter
5: Transformations and Weighting to Correct Inaccuracies
Variance
Stabilizing Transformations
Transformations to
Linearize the Model
Analytical Methods
for Selecting a Transformation
Generalized and
Weighted least Squares
Chapter
6: Diagnostics for Leverage and Influence
Importance of
Detecting Influential Observations
Leverage
Measures of
Influence (Cook’s D)
Measures of
Influence (DFFITS and DFBETAS)
Measure of Model
performance
Detecting Groups of
Influential Observations
Treatment of
Influential Observations
Chapter
7: Polynomial Regression
Polynomial Models
in One variable
Nonparametric
Regression
Polynomial Models
in Two or More variable
Orthogonal
Polynomials
Chapter
8: Indicator Variables
General Concept of
an Indicator Variable
Comments on the Use
of Indicator Variables
Regression Approach
to Analysis of Variance
Chapter
9: Variable Selection and Model Building
Model Building
Computational
Techniques for Variable Selection
Chapter
10: Multicollinearity
Sources
Effects
Diagnostics
Methods
Chapter
11: Robust Regression
Need for Robust
Regression
M-Estimators
Properties of
Robust Estimators
Other Robust
Estimators
Chapter
12: Non-Linear Regression
Linear and
Non-Linear Models
Non-Linear Least
Squares
Transformation to a
Linear Model
Parameter
Estimation
Statistical
Inference
Chapter
13: Generalized Least Squares
Logistics
Regression Models
Poisson Regression
Models
Generalized Linear
Model
Chapter
14: Other Topics in Regression Analysis
Regression Models
with Autocorrelation Errors
Effect of
Measurement Errors in the Regressors
Inverse Estimation
Bootstrapping
Classification and
Regression Trees
Neural Networks
Designed
Experiments for Regression
Chapter
15: Validation
of Regression Models
Validation
Techniques
Data from Planned
Experiments
Design
of Experiments for Business Decision Making
Seminar
Description:
This course will present tools and methodology useful in
conducting experiments that provide valid answers to questions of
interest to the experimenter.
The course will discuss an overall approach to obtaining
and analyzing experimental data, the advantages of using
structured multifactor experiments to screen for important
factors, ways of minimizing the amount of data points needed to
obtain desired information, and how to identify values of
experimental factors that optimize the value of measured
responses. Factorial
designs, fractional factorial designs, and response surface
designs will be presented. Emphasis
will be on the knowledge required for proper application of these
methods through many examples in business and quality management.
Prerequisites:
Participants in this seminar must
have achieved Six Sigma Yellow Belt
certification.
Six
Sigma Certification:
Design of Experiments for Business
Decision Making is required to sit for the Six Sigma Black Belt
examination.
Text:
Paul Mathews, “Design of Experiments with Minitab,”
American Society for Quality, ASQ
Quality Press, WI: Milwaukee, 2005. The book comes with a
supplementary CD containing Minitab files.
Statistical
Software:
Minitab 14
Seminar Outline:
Chapter
1: Graphical Presentation of Data
Types of Data
Histograms
Dot Plots
Scatter
Plots
Multi-vari
Charts
Chapter
2: Descriptive Statistics
Selection of
Samples
Measures of
Location
Measures of
Variation
The Normal
Distribution
Chapter
3: Inferential Statistics
Distribution of
Sample Means (Sigma Known)
Confidence for the
Population Mean (Sigma Known)
Hypothesis Test for
One Sample Mean (Sigma Known)
Distribution of
Sample Means (Sigma Unknown)
Hypothesis tests
fro Two means (sigma Known and Unknown)
Inferences About
One Variance
Hypothesis Tests
for Two Sample Variances
Testing for
Normality
Sample Size
Calculation
Chapter
4: DOE Language and Concepts
Definition, Scope,
and Motivation
Experiment
Defined
Identification of
Variables and Responses
Types of Variables
Types of Responses
Interactions
Types of
Experiments
Types of Models
Selection of
Variable level
Nested Variables
Covariates
Definition of Design
in Design of Experiments
Types of Designs
Randomization
Replication and
Repetition
Blocking
Confounding
Data Integrity and
Ethics
Experiment
Documentation
Why Experiments Go
Bad
Chapter
5: Experiments for One-Way Classifications
Analysis by
Comparison of All Possible Pairs Means
Graphical Approach
to ANOVE
Introduction to
ANOVA
Sum of Squares
Approach to ANOVA Calculations
Calculating Forms
for Sum of Squares
ANOVA for
Unbalanced Experiments
After ANOVA:
Comparing treatment Means
Completely
Randomized Designs
Analysis of Means
Response
Transformations
Sample Size for
One-Way ANOVA
Design
Considerations for One Way Classification Experiments
Chapter
6: Experiments for Multi-Way Classifications
Rationale for
Two-Way ANOVA
Sum of Squares
Approach to Two- Way ANOVA (One Replicate)
Interactions
Interpretation of
Two Way Experiments
Factorial Designs
Multi-Way
Classification Designs
Chapter
9: Two-Level (2k
) Factorial Designs
22
Factorial Design
23
Factorial Design
Addition of Center
Cells to the 2k Factorial Design
General procedures
for the Analysis of 2k Factorial Design
Extra and Missing
Values
Propagation of
Error
Sample Size and
Power
Chapter
10: Fractional Factorial Designs
25-1
Half Fractional Factorial Design
Other Fractional
Factorial Designs
Design Resolution
Consequences of
Confounding
Interpretation of
Fractional Factorial Designs
Plackett-Burman
Designs
Sample Size
Calculations
Design
Considerations
Chapter
11: Response Surface Designs
Terms in Quadratic
Models
2k Designs
with Centers
3k Factorial
Designs
Box-Behnken Designs
Central Composite
Designs
Comparison of
Response Surface Designs
Sample Size
Calculations
|
