Year: 2018
Type: conference-paper
Abstract: Advances in machine learning techniques in recent years were of great benefit for the detection of diseases/medical conditions in eHealth systems, but only to a limited extend.In fact, while for the detection of some diseases the data mining techniques were performing very well, they still got outperformed by medical experts in about half of the tests done.In this paper, we propose a hybrid approach, which will use a rule-based system on top of the machine learning techniques in order to optimise the results of conflict handling.The goal is to insert the knowledge from medical experts in order to optimise the results given by the classification techniques.Possible positive and negative effects will be discussed.
Authors: Gilles Neyens, Denis Zampuniéris
Institution: University of Luxembourg
Language: English
Subfield: Software
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W2801761597