| DAY ONE |
| I. |
MODELLING BASICS |
| |
A. |
Representation |
| |
B. |
Performance Criterion |
| |
C. |
Alternative Methods |
| |
D. |
Modelling Process |
| |
a. |
Variable Selection |
| |
b. |
Model Assessment |
| |
c. |
Model Validation |
| II. |
EXPLORATORY DATA ANALYSIS (EDA) |
| |
A. |
What is it, and why do it? |
| |
B. |
The Hallmarks |
| |
C. |
Stars and Profile Curves |
| |
D. |
RE-EXPRESSING |
| |
a. |
Symmetry |
| |
1) |
Ideal Shape |
| |
b. |
Straightening |
| |
1) |
Weakness of R-square |
| |
c. |
Smoothing |
| |
1) |
Critical Step |
| III. |
TREE MODELS |
| |
A. |
Primer |
| |
B. |
CHAID vs. CART vs. CLS |
| |
C. |
Many Uses |
| IV. |
RESAMPLING |
| |
A. |
Jackknife and Bootstrap |
| |
B. |
Bootstrapped Decile Analysis |
| V. |
RE-EXPRESSING MANY VARIABLES |
| |
A. |
Principal Component Analysis (PCA) |
| |
B. |
Case Studies |
| VI. |
PRINCIPAL COMPONENT ANALYSIS |
| |
A. |
Compositional Data |
| |
B. |
Relation with Factor Analysis |
| DAY TWO |
| VII. |
BINARY LOGISTIC REGRESSION |
| |
A. |
Model Specification |
| |
B. |
Linear Probability Model vs. Logit vs.
Probit |
| |
C. |
Logistic Regression Interpretation |
| |
D. |
Extensive Case Study |
| VIII. |
EDA PRODUCT AFFINITY |
| |
A. |
CHAID and PCA |
| IX. |
PCA PRODUCT AFFINITY |
| |
A. |
Case Study Illustration |
| |
B. |
PCA Segmentation |
| X. |
LOGISTIC REGRESSION FAMILY |
| |
A. |
Multinominal Logistic |
| |
a. |
Expanding the binary logistic model |
| |
b. |
Case Study |
| |
B. |
Ordinal Logistic |
| |
a. |
Specification |
| |
b. |
Key assumption |
| |
C. |
Weighted Least-Squares |
| |
a. |
Specification |
| |
b. |
Zipcode Modelling |
| XI. |
ALTERNATIVE RESPONSE MODELLING METHODS |
| |
A. |
Statistical Methods |
| |
B. |
Artificial Neural Networks |
| |
C. |
Genetic Algorithms |
| |
D. |
Comparative Evaluation |