Εφαρμογή μεθόδων στατιστικής μηχανικής μάθησης στην ανάλυση κειμένων και εικόνων υγείας
Statistical machine learning used in health image and text analysis
View/
Keywords
Machine learning ; Text mining ; Text classification ; Image classification ; HealthcareAbstract
Current thesis aims to cover a wide range of pre-processing techniques on text and image data on health sector, with the scope of improving the performance of classification algorithms. For text classification, Machine Learning algorithms such as Logistic Regression, Multinomial Naïve Bayes, and Support Vector Machines (SVM) were used and the results of them were compared to each other. The best performing algorithm was SVM with a classification accuracy of 88.98%. Deep Learning techniques were used as well, for the same task. A Multilayer Perceptron and a 1-Dimentional Convolutional Neural Network (1-D CNN) were trained and evaluated. The best performance achieved from the Multilayer Perceptron with a classification accuracy of 84.85%. Image classification task was implemented using CNNs and Transfer Learning. Specifically, a simple CNN and a pre-trained CNN known as VGG16 were used and the best performing was the VGG16 with a classification accuracy of 97.83%.
Postgraduate Studies Programme
Εφαρμοσμένη ΣτατιστικήDepartment
Σχολή Χρηματοοικονομικής και Στατιστικής. Τμήμα Στατιστικής και Ασφαλιστικής ΕπιστήμηςNumber of pages
142Language
EnglishCollections
Except where otherwise noted, this item's license is described as
Αναφορά Δημιουργού-Μη Εμπορική Χρήση-Όχι Παράγωγα Έργα 3.0 Ελλάδα
Αναφορά Δημιουργού-Μη Εμπορική Χρήση-Όχι Παράγωγα Έργα 3.0 Ελλάδα