Syllabus
The following topics will be presented:<br>
1. Typology of multidimensional methods. Basic descriptive methods and graphs. Smooth introduction to multidimensional geometry. <br>
2. Principal Component Analysis}: geometry, interpretation and usage.<br>
3. Factor Analysis: theoretical assumptions, geometry, implications, description, interpretation and prediction. Relation to PCA.<br>
4. Cluster Analysis.<br>
5. Discriminant analysis. Linear, Fisher's, quadratic ... Introduction to classification.<br>
6. Classification and Regression Trees (CART). Slight introduction to other (non-linear) methods (neural networks, SVM). Measurement of classifiers' quality.<br>
7. Regression and Generalized Linear Models.<br>
8. Logistic regression.<br>
9. Log-lineár regression models and analysis of contingency tables.
Annotation
The course is an introduction to a broad spectrum of multidimensional methods of statistical analysis. The techniques useful potentially in sociology will be focused as well as their principal properties and limitations.
Repetitive enrollment is allowed.
