Crowdsourced healthcare knowledge creation using patients’ health experience-ontologies

In this research, we developed CHEKC framework for creation and integration of crowdsourced healthcare knowledge using experience-ontologies. The purpose is to provide patients’ healthcare information which contains similar healthcare experiences including conditions and symptoms and integrates the features and relations in the particular patients’ data according to users’ queries. To do this, we developed three modules and ontologies. The modules are Crowdsourced Health Data Manipulation Module (CHMM), Health Ontology-based Relevant Patient Finding Module (HRFM), and Ontology-guided Healthcare Knowledge Integration Module (OKIM). CHMM is developed to transform patients’ data to structured cases with problem-solution. The cases are stored in CHEKC Patient Ontology. HRFM is developed to find relevant cases according to the user’s query using CHEKC Upper Ontology. To do this, ensemble semantic similarity is proposed using semantic similarity and fuzzy C-means clustering and the relevant cases are stored in Interim Ontology. OKIM is developed for the integration of the relevant cases using SWRL rule-base. However, it is not guaranteed to find suitable rules and generate necessary knowledge from the rule-base. To relieve the problem, ontology-guided knowledge integration is proposed, which supports the inferring relations among classes in CHEKC Interim Ontology. CHEKC framework provides the integrated healthcare information and knowledge which are generated through the illustrated processes using the selected similar healthcare cases with users’ query. In particular, the cases are constructed by crowdsourcing on healthcare-featured social media and are based on patients’ healthcare experiences from the perspectives of patients. Through the conducting of two experiments, we proved the effectiveness of CHEKC framework. The conducted experiments proved the efficiency of CHEKC framework by the reduction in search volumes and the improvement in accuracy of query results.