Connectivity and Pattern Recognition

Recommended Prerequisites

The student is expected to be reasonably comfortable (but not necessarily an expert) with certain foundational topics such as linear algebra, probability and statistics, functional analysis and programming.

Teaching Methods

Theoretical classes with detailed presentation, using audiovisual means, of the concepts, principles and fundamental theories and with the introduction of several examples to illustrate the practical interest of the subject and exemplify its application to real cases.

Learning Outcomes

To provide the student with an in-depth contact with methodologies and techniques  of pattern recognition for understanding:

- the essential concepts of a pattern recognition system;

- the nature and inherent difficulties of the pattern classification problem;

- concepts such as trade-offs and appropriateness of classification techniques such as Bayesian, maximum-likelihood, nonparametric, linear discriminant, principal components;

- advanced concepts such as methods for dimension space reduction namely manifold learning algorithms (e.g., Isomap and Local Linear Embedding (LLE);

- deal with case-studies to develop their skills in addressing decision and pattern recognition problems;

- provide the students with the main methodologies for the development of solutions for the automatic analysis of bio signals (e.g.,time series). To apply algorithms in the classification and diagnosis of bio signals in the context of clinical decision support;

- text classification applications.

Work Placement(s)

No

Syllabus

Module I -Pattern Recognition: Concepts, Methods and Applications
1. Pattern Similarity and Pattern Recognition Tasks
2. Classes Patterns and Features
3. Pattern Recognition Approaches
4. Data Clustering
5. Statistical Classification
6. Dimensionality Reduction
Module II - Biomedical Applications
1. Introduction: main characteristics of bio signals;
2. Biosignal analysis: time domain and frequency analysis; sampling
3. Electrocardiogram: cardiovascular system; main arrhythmias;
4. Electrocardiogram: algorithms for the ECG segmentation; feature extraction for the arrhythmias’ classification; heart rate variability analysis.
Module III - Text and Web Mining Applications

Head Lecturer(s)

Bernardete Martins Ribeiro

Assessment Methods

Assessment
Exam: 50.0%
Paper and presentation of the work based on a project: 50.0%

Bibliography

1 . Pattern Recognition: Concepts, Methods and Applications - Marques de Sá, J.P. (Springer-Verlag, 2001)
2. Pattern Classification, Duda, R. O., Hart, P.E., and Stork, D.G. ( ISBN: 0-471-05669-3. 2nd ed. Wiley Interscience, 2001)
3. Pattern Recognition and Machine Learning, - Bishop, C.M. (Springer Verlag, 2006)
4. Reconhecimento de Padrões: Métodos Estatísticos e Neuronais - Jorge Salvador Marques (ISBN: 972-8469-08-X Ano: 2005 2ª Edição http://istpress.ist.utl.pt/lrecpad.html )
5. Handbook of Medical Informatics; Van Bemmel, J.H. & Musen, M.A. (Eds.); Springer-Verlag, 1997, ISBN 3-450-63351.
6. Bioelectrical Signal Processing in Cardiac and Neurological; Leif Sornmo, Pablo Laguna; Elsevier Science & Technology; 2005; ISBN-13: 9780124375529.
7. Introduction to Biomedical Engineering; Enderle  & Bronzino; Academic Press; 2011;
ISBN: 978012374979 .
8. J. Solka, “Text Data Mining: Theory and Methods”, Statistic Surveys, 2008