Analysis children with disabilities self-care problems based on selected data mining techniques

Abstract

Children’s disabilities is a very important task. Lack of appropriate diagnosis and care increases children’s difficulties in adjusting to disability. Children with physical and motor disabilities have many problems in their lives related to basic daily activities. Therefore, it is very important to properly classify self-care problems. Such classification requires appropriately trained people and time. Data mining techniques can be helpful to support such diagnosis. Classification and selection of appropriate features can significantly support the work of diagnosticians and therapists. In this paper we attempted to build classification models based on all available instances as well as on instances obtained by feature selection. The most accurate model obtained, correctly classifies more than 85% of children with self-care problems, using only 19 characteristic features.