Towards Rules Induction Optimization for Cerebral Palsy Diagnosis: An Empirical Study

According to the World Health Organization, there are about 15% of universe population have some forms of disability. One of these disabilities is Cerebral Palsy CP which is a permanent motor disorder that appears in early age causing mental and physical disabilities. Based on WHO, between 1.5 to 2.8 cases per 1000 birth have CP Disability in the world. In Gaza Strip, 2.6% of population has some forms of disability. The children under 18 years with CP disability represent 3.1% of disability population in Gaza Strip. Based on domain experts' opinion, there are very little studies that target disability in Gaza Strip especially Cerebral Palsy. In this paper we propose a new approach for rules induction and optimization for Cerebral Palsy diagnosis based on symptoms that appear on the children aged from 1 day to 3 years. We used knowledge-based survey technique to extract tacit knowledge related to disability symptoms that indicate Cerebral Palsy. These symptoms have been weighted and combined with a set of rules that are manually built by domain experts. This combination optimizes diagnosis rules of Cerebral Palsy in early ages. Our proposed approach succeeded to diagnose 90% of a set of pre-diagnosed cases. The proposed approach of rule induction optimization of CP is an important phase toward ontology engineering for CP diagnosis.