University of Thessaly (UTH)
Design and implementation of machine learning algorithms in big biomedical data
Abstract
dc:descriptionMachine Learning (ML) is a critical tool for Big Biomedical data analysis based on the fundamental principle that data quality determines the efficacy of analytical results. This investigation emphasizes the importance of meticulous data collection and preprocessing, focusing on precision, to support the dependability of ML applications. It addresses the critical issues of missing information, anomalies, duplications, and boundary cases, all critical for priming datasets for comprehensive analysis. The emphasis then shifts to the art and science of data preprocessing, with aggregation, sampling, dimensionality reduction, feature selection, and engineering highlighted as key techniques for transforming raw data into structured formats suitable for advanced ML algorithms. These preliminary steps are critical for leveraging ensemble learning, which emerges as a strong strategy for improving model performance. Ensemble learning, which combines multiple predictive models, is proposed as a way to improve accuracy and robustness, particularly in the context of complex Big Biomedical datasets. This thesis investigates neural network performance optimization and presents projection-based ensemble classifiers with neural network voting. The practical application of these concepts is demonstrated through the introduction of a hierarchical divisive clustering toolbox, exemplifying the effective use of projection-based ensemble methods for sophisticated clustering and classification tasks. The methodologies and tools presented provide a comprehensive strategy for harnessing the complexities of biomedical big data, with the goal of improving analytical capabilities in this domain.
Degree
thesis:*- Grantor dc:publisher
- University of Thessaly (UTH)
- Year dc:date
- 2023
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Anagnostou, Panagiotis
- Αναγνώστου, Παναγιώτης
Subjects
dc:subject × 38- Μηχανική μάθηση
- Μεγάλα δεδομένα
- Βιοπληροφορική
- Υψηλή διαστατικότητα
- Αντικατάσταση κενών
- Βελτίωση πρόβλεψης
- Μέθοδοι συνόλου
- Αλγόριθμοι ομαδοποίησης
- Αλγόριθμοι ταξινόμησης
- Δεδομένα ορίων
- Νευρωνικά δίκτυα
- Υπερεπίπεδα ελάχιστης πυκνότητας
- Προσεγγιστικός πλησιέστερος γείτονας
- Έργο ATHLOS
- Υγειονομική περίθαλψη
- Ανάλυση δεδομένων
- Machine learning
- Big data
- Bioinformatics
- High dimensionality
- Imputation
- Prediction enhancement
- Ensemble methods
- Clustering algoriths
- Classification algorithms
- Boundary data
- Neural networks
- Minimum density hyperplanes
- Approximate nearest neighbor
- ATHLOS Project
- Healthcare
- Data analysis
- Επιστήμη Ηλεκτρονικών Υπολογιστών και Πληροφορική
- Φυσικές Επιστήμες
- Τεχνητή νοημοσύνη
- Computer and Information Sciences
- Natural Sciences
- Artificial Intelligence
Rights
- Language dc:language
- eng
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/56040