Πρόβλεψη θνησιμότητας από καρδιαγγειακά νοσήματα με τεχνικές μηχανικής μάθησης : ανάλυση χρονοσειρών και εφαρμογή LSTM
Cardiovascular mortality prediction using machine learning techniques : time series analysis and LSTM application
View/
Keywords
Καρδιοαγγειακά νοσήματα ; Θνησιμότητα ; Μηχανική μάθηση ; LSTM ; ARIMAAbstract
This thesis investigates cardiovascular disease (CVD) mortality and develops a machine
learning predictive model based on Long Short-Term Memory (LSTM) for estimating
CVD trends up to 2035. Using historical mortality and population data, the analysis
reveals a rising trend in CVD mortality, with predictions suggesting that global CVD
deaths may exceed 243 per 100,000 population by 2035. In Greece, the mortality rate is
expected to reach 482.4 per 100,000, reinforcing the need for effective public health
interventions.
The comparison of different forecasting models highlighted the superiority of LSTM
over traditional statistical approaches such as ARIMA, achieving lower prediction error
and higher accuracy. Additionally, the study emphasized the impact of socioeconomic
and environmental factors on CVD progression, advocating for the integration of more
variables
into
predictive
models.
The research concludes that AI applications in CVD forecasting can be a useful tool for
public health policy-making, enabling targeted preventative interventions. Future
improvements include developing hybrid LSTM-Transformer models and gathering
more comprehensive data to enhance prediction accuracy and reduce cardiovascular
mortality.
Postgraduate Studies Programme
Ψηφιακά Συστήματα και ΥπηρεσίεςDepartment
Σχολή Τεχνολογιών Πληροφορικής και Επικοινωνιών. Τμήμα Ψηφιακών ΣυστημάτωνNumber of pages
74Language
GreekCollections
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