Please use this identifier to cite or link to this item:

http://hdl.handle.net/10609/82006
Title: Análisis comparativo de diferentes métodos de agrupación para el tratamiento de datos de expresión genética
Author: Gómez Sánchez, Juan Alberto
Director: Rebrij, Romina
Tutor: Merino Arranz, David
Others: Universitat Oberta de Catalunya
Keywords: gene expression
machine learning
clustering
Issue Date: Jun-2018
Publisher: Universitat Oberta de Catalunya
Abstract: The aim of this project is to obtain a method that allow us to compare how some of the most common clustering algoritms works when they are applied over a genetic expression matrix. The desire to select the most efficient algorithms arises from the progressive increase of this data type volume and the huge requirements that it needs. To achieve the objetives it has been developed all the diferents chosen algoritms using it over a common data set which consist in samples of mammary tissue that are divided by his celular type (lobular and ductal) and by his state (normal and tumoral), and its results has been compared in function of diverse criteria like internality, stability and biological variation. Results of how this algorithms group the data has been obtained along with a comparative about the process. The results point to a better control of this data by the hierarchical algorithms, especially the DIANA algorithm. Finally, it is concluded that this type of algorithms doesn't works well with this type of data, because it didn´t get a good divide on the data set generally, being the division between tumoral and normal cells the only what can be rating as positive.
Language: Spanish
URI: http://hdl.handle.net/10609/82006
Appears in Collections:Bachelor thesis, research projects, etc.

Share:
Export:
Files in This Item:
File Description SizeFormat 
jgomezsanchezTFM0618memoria.pdfMemoria del TFM879.47 kBAdobe PDFView/Open
jgomezsanchezTFM0618código.pdfCódigo en formato Markdown357.26 kBAdobe PDFView/Open

This item is licensed under a Creative Commons License Creative Commons