A carregar...
Publicação
Unsupervised feature extraction with autoencoder : for the representation of parkinson´s disease patients
| Nome: | Descrição: | Tamanho: | Formato: | |
|---|---|---|---|---|
| 3.76 MB | Adobe PDF |
Autores
Orientador(es)
Resumo(s)
Data representation is one of the fundamental concepts in machine learning. An appropriate representation is found by discovering a structure and automatic detection of patterns in data. In many domains, representation or feature learning is a critical step in improving the performance of machine learning algorithms due to the multidimensionality of data that feeds the model. Some tasks may have different perspectives and approaches depending on how data is represented. In recent years, deep artificial neural networks have provided better solutions to several pattern recognition problems and classification tasks. Deep architectures have also shown their effectiveness in capturing latent features for data representation.
In this document, autoencoders will be examined to obtain the representation of Parkinson's disease patients and compared with conventional representation learning algorithms. The results will show whether the proposed method of feature selection leads to the desired accuracy for predicting the severity of Parkinson’s disease.
Descrição
Dissertation presented as partial requirement for obtaining the Master’s degree in Information Management, with a specialization in Knowledge Management and Business Intelligence
Palavras-chave
Autoencoder Representation Learning Feature Extraction Unsupervised Learning Deep Learning