Propostas de Estágio 2012/2013

DEI - FCTUC
Gerado a 2024-11-24 10:08:16 (Europe/Lisbon).
Voltar

Titulo Estágio

Algorithms for prediction of cardiovascular events

Área Tecnológica

Informática Médica

Local do Estágio

DEI

Enquadramento

In the context of telemonitoring, the identification of cardiovascular events are usually based on models that uses continuous update of measurements, parameters and symptoms, collected during daily home monitoring process. Basically, these models assume that these conditions can be characterized based on biosignals and on their dynamic evolution. Examples of these are hypertension, myocardial ischemia, arrhythmias, pulmonary edema, etc., which are themselves defined through literature or by clinical expertise

Objetivo

This work aims the development of a predictive strategy able to estimate future events with relevant impact in the cardiovascular status. The predictor strategy to be developed is supported by the historic values of daily measurements. This strategy can be used to predict individual variables or in a multi-parametric scheme to predict simultaneously several variables.
Two key aspects will be addressed:
i) Development of techniques for similarity detection in bio-signal time series, which permit to detect specific events. In particular, methods that exploit the time-frequency characteristics of the signals will be investigated.
ii) Development of prediction methods to be applied in the trend analysis of the bio-signals and detected events. For this purpose, neural networks or fuzzy sytems are promising solutions.
The result of these predictions can be useful in simple future events detection, in alerts generation when integrated in clinical decision support systems, as well as in short-term CV risk models to derive a dynamic CV risk scheme.
Moreover, these algorithms may help in finding groups of patients with similar clinical behavior, thus in the identification of temporal patterns that may be suggestive of different clinical events.

Plano de Trabalhos - Semestre 1

1. Clinical context
2. State of the art: similarity and prediction techniques
3. Data base analysis
4. Dissertation intermediate documentation

Plano de Trabalhos - Semestre 2

1. Development of algorithms for the prediction of CV events
2. Validation
3. Integration into the clinical framework
4. Writing of dissertation

Condições

A realizar no DEI

Orientador

Jorge Henriques
jh@dei.uc.pt 📩