Publication in the Diário da República: Despacho n.º 12419/2016 - 14/10/2016
6 ECTS; 3º Ano, 1º Semestre, 30,0 T + 45,0 TP + 5,0 OT , Cód. 814320.
Lecturer
- Sandra Maria Gonçalves Vilas Boas Jardim (2)
(1) Docente Responsável
(2) Docente que lecciona
Prerequisites
The course builds on the knowledge acquired in Calculus, Discrete Mathematics, Programming and Algorithmics and Programming Languages.
Objectives
1. Basic principles, mathematical foundations and technical applications of machine learning.
2. The strengths and weaknesses of different algorithms for different domains of application.
3. The overlearning phenomenon.
Program
I. Introduction
II. Parametric regression
III. Fundamentals
IV. MV method and regression models
V. Classification concepts
VI. Bayes Theory
VII. Linear discriminant classification
VIII. Gradient descent methods
IX. Classification with logistic discriminants and neuronal networks
X. Non-supervised methods
XI. Decision trees
Evaluation Methodology
Written test worth 40% of the final mark.
Battery of problem solving tasks worth 60% of the final mark.
Bibliography
- Jensen, F. e Nielsen, T. (2007). Bayesian Networks and Decision Graphs (Information Science and Statistics). (Vol. 1). (pp. 1-448). USA: Springer
- Marques, J. (2005). Reconhecimento de Padrões - Métodos Estatísticos e Neuronais. (Vol. 1). (pp. 1-284). Lisboa: IST Press
- Stork, D. e Hart, P. e O. Duda, R. (2000). Pattern Classification. (Vol. 1). (pp. 1-635). USA: Wiley-Interscience
Teaching Method
Classes include topic presentation and practical cases. Key topics are supported by exercises and computer-based practical coursework.
Software used in class
