Unit:
Library of the School of Health SciencesDepartment of Nursing
Dissertation committee:
Ιωάννης Μαντάς, Καθηγητής, Τμήμα Νοσηλευτικής, ΕΚΠΑ
Φλώρα Μαλαματένιου, Καθηγήτρια, Τμήμα Ψηφιακών Συστημάτων, ΠΑΠΕΙ
Δάφνη Καϊτελίδου, Αναπληρώτρια Καθηγήτρια, Τμήμα Νοσηλευτικής, ΕΚΠΑ
Γεώργιος Φιλντίσης, Καθηγητής, Τμήμα Νοσηλευτικής, ΕΚΠΑ
Αθηνά Καλοκαιρινού, Καθηγήτρια, Τμήμα Νοσηλευτικής, ΕΚΠΑ
Παναγιώτης Μπαμίδης, Καθηγητής, Τμήμα Ιατρικής, ΑΠΘ
Μαριάννα Διομήδους, Καθηγήτρια, Τμήμα Νοσηλευτικής, ΕΚΠΑ
Original Title:
Εξόρυξη γνώσης από αρχεία μεγάλου όγκου δεδομένων υγείας -big data-με χρήση υπολογιστικών αλγορίθμων ανάλυσης-health analytics
Translated title:
Data mining of large healthcare data -big bata-using analytic algorithms-health analytics
Summary:
The biggest challenge of modern computing is undoubtedly the efficient storage and retrieval of very large amounts of data. This need made its appearance in recent years due to the explosion of data that is observed on the Internet and is becoming more important because of the very wide range of information we can learn. The fields of healthcare and medical data are constantly evolving. The use of BigData in healthcare offers valuable information as they are capable for effective storage, data processing, sql queries and health data analysis.
The aim of the current PhD thesis is the study and the implementation of data mining techniques using Big Data, in the field of healthcare. Also, the current PhD thesis is to build and compare classification techniques for cardiovascular diseases. Finally, the research aim of this study is to investigate the perceptions of the Health Informatics Scientists and Health Professionals about the Big Data Technology in Healthcare.
On this Doctoral Thesis, a literature review was conducted in order to record the current status of Big Data in Healthcare. This review records the definition of Big Data, the attributes, the advantages and disadvantages of Big Data in Healthcare. Furthermore the review presents the implementation and storage mechanisms of Big Data. Also, the review records software for analysis and processing Big Data, Big Data programming languages. The use of Big Data in Europe and in the world is also presented. Finally, the basic principles of the GDPR are presented and it’s correlation to Big Data in the field of Healthcare. Two empirical studies were also conducted.
The research aim of the first study is to investigate the perceptions of Health Information scientists about the Big Data Technology in Healthcare. A questionnaire was developed in order to measure the aforementioned dimensions. The current study reveals a rather positive attitude toward the usage of Big Data in the Healthcare domain.
The research aim of the second study is to investigate the perceptions of Health Professionals about the Big Data Technology in Healthcare. A questionnaire was developed in order to measure the aforementioned dimensions. The current study did not give sufficient results as the respondents showed a positive attitude towards Big Data although they do not know much about this technology.
The last part of the thesis refers to the development of data mining techniques for the prediction of cardiovascular diseases. The methods used are: Logistic Regression, Naive Bayes Classifier, Decision Trees, K Nearest Neighbors Algorithm, SVM (Support Vector Machine) Algorithm and Random Forest. The development included all stages of data preprocessing while specific metrics were used to measure the performance of the classifiers. Finally, the performance of the classifiers was improved using cross-validation, while the problem of class imbalance was solved using the SMOTE method.
Main subject category:
Health Sciences
Keywords:
Empirical study, Big data, Health professionals, Data mining, Classification techniques
Number of references:
130
