| Name: | Description: | Size: | Format: | |
|---|---|---|---|---|
| 5.82 MB | Adobe PDF |
Authors
Advisor(s)
Abstract(s)
Atualmente, o setor da saúde é um dos setores que apresenta um crescimento mais rápido. Os dados crescem exponencialmente todos os dias e à medida que a quantidade de informação aumenta, novas formas de interagir e extrair significado vão surgindo. Machine Learning é uma área da Inteligência Artificial que estuda como solucionar problemas complexos e intuitivos, apresentando potencial para ser a solução que permitirá reduzir o custo crescente dos cuidados de saúde e, auxiliar os médicos no tratamento dos seus pacientes, de forma mais rápida e eficaz. A presente dissertação de mestrado tem como objetivo estudar os perfis metabólicos de aminoácidos e acilcarnitinas em doentes com cancro do reto localmente avançado e a forma como estes perfis influenciam a biologia tumoral e a resposta ao tratamento. Pretende-se analisar os respetivos perfis para a realização de duas tarefas de aprendizagem não supervisionada: Deteção de anomalias e Clustering. No decorrer deste trabalho, serão testados diferentes algoritmos e técnicas de tratamento de dados para serem avaliados por métricas intrínsecas para determinar o nível de coesão, separação e semelhança nos resultados de grupos obtidos, seguido pelas respetivas conclusões. Os melhores algoritmos serão utilizados num sistema de decisão que ajude de forma mais eficiente os profissionais de saúde na escolha de tratamento adequado para cada um dos futuros pacientes.
Currently, the healthcare sector is one of the fastest growing sectors. Data exponentially takes into account every day and as the amount of information increases, new ways of interacting and extracting meaning are emerging. Machine Learning is an area of Artificial Intelligence that studies how to solve complex and intuitive problems, presenting the potential to be the solution that will reduce the growing cost of health care and help doctors in treating their patients more quickly and effectively. This master's thesis aims to study the metabolic profiles of amino acids and acylcarnitines in patients with colorectal cancer and how these profiles influence tumor biology and response to treatment. It is intended to analyze the respective profiles for carrying out two unsupervised learning tasks: Anomaly detection and Clustering. In the course of this work, different algorithms and data processing techniques will be tested to be evaluated by intrinsic metrics to determine the level of cohesion, separation and similarity in the obtained group results, followed by the respective understanding. The best algorithms will be used in a decision system that will more efficiently help health professionals in choosing the right treatment for each of their future patients.
Currently, the healthcare sector is one of the fastest growing sectors. Data exponentially takes into account every day and as the amount of information increases, new ways of interacting and extracting meaning are emerging. Machine Learning is an area of Artificial Intelligence that studies how to solve complex and intuitive problems, presenting the potential to be the solution that will reduce the growing cost of health care and help doctors in treating their patients more quickly and effectively. This master's thesis aims to study the metabolic profiles of amino acids and acylcarnitines in patients with colorectal cancer and how these profiles influence tumor biology and response to treatment. It is intended to analyze the respective profiles for carrying out two unsupervised learning tasks: Anomaly detection and Clustering. In the course of this work, different algorithms and data processing techniques will be tested to be evaluated by intrinsic metrics to determine the level of cohesion, separation and similarity in the obtained group results, followed by the respective understanding. The best algorithms will be used in a decision system that will more efficiently help health professionals in choosing the right treatment for each of their future patients.
Description
Keywords
Machine Learning Aprendizagem não supervisionada Colorretal Aminoácidos Acilcarnitinas Unsupervised learning Coloretal Amino acids Acylcarnitines