Browsing by Author "Yesigat, Simeneh"

Now showing 1 - 1 of 1
  • No Thumbnail Available
    Item
    Application Of A Hybrid Recommender System For Ethiopian Local Food Selection To Diabetes Type 2 Patients
    (Addis Ababa University, 2015-10-03) Yesigat, Simeneh; Jemaneh, Mr. Getachew
    Diabetes is becoming one of the rapidly increasing non-communicable diseases and an important public health problem all over the world. Ethiopia is at a risk of increased diabetes incidence. The number of deaths attributed to diabetes reached over 21,000 in 2007. Among this number, Type-2 diabetes constitutes about 85 to 95% of all diabetes. Dietary management is considered to be one of the cornerstones of diabetes care. Good diabetes management is a balance between healthy eating, exercise and medication. The problem, however, is that most diabetic patients have difficulty of identifying the recommended quality and quantity of food that they have to eat in order to control their blood glucose. The aim of the study is to assess the applicability of a hybrid filtering approach to create a personalized food recommender system that can promote healthier eating habits. So that, the patients can safely chooses a healthier food that can promote their self-dietary management. To achieve this goal, the study attempts to design and develop a prototype hybrid recommender system which is a combination of both collaborative, content base and demographic filtering methods with a mixed, cascade and switching hybrid strategies. The knowledge is acquired using unstructured interviews from domain expert and relevant documents analysis method is also applied to capture explicit knowledge. The data about Ethiopian foods is collected from the food composition table from ENHRI booklet. The prototype is developed by using SQL Server 2008 database with a visual Basic programming language. Moreover, in testing and evaluating the prototype system five volunteer patients were involved. The experiment conducted in five iterations for each user. So the result shows that, the recommender system has a good performance and achieved the desired goal. But further research is needed to be conducted of exploring different options of hybrid strategies.