Εφαρμογή μοντέλων πρόβλεψης με χρήση του εργαλείου της R για την πρόβλεψη εμφάνισης καρδιακών προβλημάτων σε ασθενείς
A predictive analysis on heart diseases using machine learning techniques with the R tool
View/ Open
Keywords
Μηχανική μάθηση ; Καρδιακές παθήσεις ; Αλγόριθμοι ; Εξόρυξη δεδομένων ; Ανάλυση δεδομένων ; Data mining ; Data analysis ; Data exploration ; Heart disease ; Διερεύνηση δεδομένωνAbstract
According to statistical analysis provided from the organization of CDC (Centers for Disease Control and Prevention), which is the most important organization of public health in the United States of America heart disease is one of, if not the most critical causes of death not only in America but in the whole world. Specifically more than 600.000 of people are affected and end up losing their life from some kind of heart disease. That is about 7 percent and almost the ¼ of the total annual deaths in America and the rest of the world.
The terminology given as heart disease refers to different types of conditions with the most crucial and common amongst them being the Coronary Artery Disease. This type of disease is the most usual and commonly found in the samples of heart disease patients. In particular, in the year 2017 there was about 365 thousands of deaths caused by CAD whereas around 7 percent of the people over 20 years old have this specific disease.
The main way of dealing with that kind of conditions are prediction and forecasting of them via the symptoms or other relevant characteristics so that it is avoided. In case it is not timely predicted the only countermeasures are the change of habits and way of life.
As easy as is to see then this type of disease is crucial and most important to be predicted via the symptoms and the characteristics which are relevant, so that more lives are saved.
Regarding forecasting we are referencing the prediction of the possibility of people having CAD or in general heart diseases in correlation with other characteristics. In today’s world and reality, where data are in the center of processes in every aspect of the world, the most reliable and accountable method for prediction is data mining using machine learning algorithms. Through machine learning we have access in a great variety of algorithms which provide accuracy and flexibility for analyzing data and predicting possible outcomes.
In the aforementioned paper we will try using methods of data mining and exploratory analysis and by visualizing the data to apprehend the correlation of the variables with the appearance of heart disease, we will understand the connection of other symptoms with them and how accurate we can predict a heart disease through the correlated variables.