Processament de llenguatge natural amb Word2Vec i màquines de suport vectorial

Abstract

[en] In this project, the theoretical foundations of two machine learning methods are studied in detail to understand their methodology and learning process. Word2Vec consists of a neural network capable of processing text data as numerical vectors that attempt to reflect the semantic and grammatical relationships between words. Support Vector Machines construct a hyperplane or set of hyperplanes in a high dimensional (or even infinite) space that can be used in classification or regression problems. Finally, these two models will be combined in an algorithm that will allow us to carry out a multiclassification task on text data, which will involve detecting possible metastasis in cancer patients from clinical annotations.