Article Preview
Top1. Introduction
This study focuses on liver fibrosis disease due to its importance, where it is result in increasing death-rate in Egypt. Viral Hepatitis C is a chronic liver disease that causes fibrosis then subsequently liver cirrhosis. This fibrosis makes changes to the architecture of the liver and disrupts its normal functions, leading to liver cancer on the long term (Suk, 2015). Liver fibrosis consists of five stages: : negative fibrosis, : mild fibrosis, : significant fibrosis, : cirrhosis, and : significant cirrhosis (Badria, 2007). Since this disease is a very critical issue for chronic HCV patients, the investigation of all the related features ‘demographic, laboratory test, clinical data and risk factors diseases’ was needed to accurate prediction of the diagnosis, thus, save money and life, and help experts to take an appropriate decision for treatment (Keltch, 2014). Poor diagnosis will lead to significant liver fibrosis, cirrhosis or to the liver’s end-stage failure (Suk, 2015). The technology of health information systems and medical devices in clinical domains increased rapidly in biomedical engineering, since such technology produces a vast amount of medical data (Kolias, 2014). However, the storage and retrieval of this produced data is difficult, and it also contains hidden, unexploited knowledge. Therefore, to overcome these problems, the scientific community suggests the usage of semantic web services (Kolias, 2014), which aims at creating ontology based on machine-processable web contents (Tho & Hui, 2006). Clinical decision support systems CDSSs are considered as intelligent information systems, which based on knowledge base formalization integrated with patient clinical data to assist physicians in their practices (Galopin, 2015). The core role of ontologies is the knowledge formalization. Therefore, it is required a set of mathematical processes and relations to construct an ontology (Maedche & Staab, 2001). Since the construction of ontology is a complicated process, there were multiple methods and tools to create patterns from different data sources such as the development of free texts (Maedche & Staab, 2000), dictionaries (Hearst, 1992) and data models (Carvalho, 2017). Using the aforementioned resources, since the data models have a widespread use in information systems, the data conceptual modeling was attracted in data modeling and management. Many formats have been used to construct ontologies such as Unified Modeling Language (UML) (Zhang & Ma, 2013), relational models (Ma, 2011), and object-oriented data models (zhang, 2010). Yet, Classical data models and the methods of building ontology are not enough to deal with inaccurate information common in the real world, especially in the medical domain (Maedche & Staab, 2001). For example, if an old-patient case has a chronic infection of Viral Hepatitis C and has high serum bilirubin (SB) laboratory test, this patient has almost Jaundice with high degree and has the majority diagnosis is liver cirrhosis. Hence, the problem is how to deal with fuzzy information with ontology and data model as extended entity relationship (EER) model.
In the daily life, we often use terms such as (almost, majority) that contain certain ambiguous or uncertainties that traditional system could not understand. Fuzzy logic (Zadeh, 1965), with its concepts make the intelligent system closer to human methods. Fuzzy systems were extensively studied to manage the storage of fuzzy data and fuzzy query capability (Galindo, 2004). Galindo et al., (2004) dealt with the EER model as a tool for dealing with fuzzy data, where they defined the fuzzy EER aspects related to the aggregation and specialization. Furthermore, Ma (2009) proposed the mapping of EER schema from the relational database schema.