Teaching construction classification rules with aplications in social sciences

Congreso; 7th. International Conference on Teaching Statistics; 2006

International Association for Statistical Education (IASE)

The supervised classification or supervised pattern recognition is a method to solve decision problems in Social Sciences. It is organized on the basis of specifics sets of predictor variables and the existence of a priori known classes. The main objective is to construct classification rules in order to predict class membership of new objects on the basis of a training sample. Nowadays, with the availability and efficacy of modern computing power, many new advances in this framework have been achieved in Statistics and the computer sciences. In this session, we discuss different methods and illustrate the results reached along several applications, covering the following topics: Parametric Discrimination, Nonparametric Discrimination, Logistic Discrimination and Neural Networks, Recursive Partitioning and Estimation of Error Rates.