Diogo Machado1, Vítor Santos Costa2, Pedro Brandão1, Inês Dutra2
1Instituto de Telecomunicações, Faculty of Sciences, University of Porto, 2CRACS, INESC-TEC, Faculty of Sciences, University of Porto
Diabetes management is a complex and a sensible problem as each diabetic is a unique case with singular needs. The optimal solution would be a constant monitoring of the diabetic’s values and, automatically, act accordingly. However, these systems are not yet available.
We propose an approach that guides the user and analyses the data gathered to give individual advice. By using data mining algorithms and methods, it uncovers hidden behaviour patterns that may lead to crisis situations. These patterns are then transformed into logical rules, able to trigger in a particular context, and advise the user, this way avoiding possible crisis.
During the data gathering phase, when the number of records is not enough to attain useful conclusions and generate individual rules, a base set of logical rules, defined from medical protocols, directives and/or advise, is responsible for advising and guiding the user on a plethora of situations. The proposed system will accompany the user at start with generic advice and, with constant learning, advise the user more specifically.
By analyzing data from different users, it is possible to determine common patterns within the diabetic population and create generic rules, relevant for these specific groups. These rules’ main goal is to guide and advise, but can also be used to describe user’s diabetes characteristics.
We believe our approach may be a relevant addiction to the monitorization process of diabetic patients and as a tool for medical experts to trace patterns and as an auxiliary information source while prescribing a treatment.
keywords: Diabetes, Rule Based Systems, Data Mining
Presentation: MyDiabetes – Rule Based Advice System