Ανάπτυξη μοντέλων πρώιμης διάγνωσης της ασθένειας Alzheimer's, με χρήση κλινικών και γενετικών δεδομένων
Date Issued
June 16, 2023
Type
Πτυχιακή Εργασία
Abstract
The current thesis’s goal is the creation of a web application which implements a machine learning model, capable of diagnosing the Alzheimer’s disease in it’s early stages. A series of events and steps were crucial for the creation of the web application. A dataset was obtained by Kaggle, one of the largest data science communities. The dataset had 820 entries and each one of them had 31 clinical and genetic features. A raw and unedited dataset is never a good idea in machine learning and for that, some actions took place. These actions include some pre-editing and visualization, like checking the correctness of the attributes, adding and removing features, and visualizing the correlation of the target feature with the other ones. After the pre-editing stage, the dataset was ready to be introduced to the algorithms, in order for the testing / training phase to take place. After a series of experiments and trials with the algorithms and their parameters, the best option was using the Random Forest Classifier with the FCBS. The final stage was creating the web application in a way for the user to be able to enter the values of clinical and genetic data, so that he/she can receive a diagnosis. The web application was written in Python and JavaScript and in order to achieve a friendlier user-interface, some CSS features were added.
Subjects