Contents I. Introduction
Diabetes is a complex and potentially debilitating chronic disease [1]. It has a widespread impact on people's health and economic impact on national healthcare systems worldwide [2]. This is a situation, where hyperglycemia can occur due to insulin secretion or different activity issues. In case of hyperglycemia, the diabetic people experience organ damage [3]. Organs affected through this disease are heart, eyes, kidney, skin, foot, and nerves system etc., and these organs can sustain long-term harm or ultimate failure [4]. A study says that these organ failures, which could have been avoided or slowed down, may cause nearly 70% of diabetes patients to pass away in the early phase of their lives. Due to this diabetes requires constant medical care and to educate patients on self-care management to avoid fatal complications and reduce the chances of long-term effects [5], [6]. Big data analytics, semantic web modelling and data science can be applied to improve the quality of diagnosis and treatment process of medical services for early-stage detection of any disease [7]. As the diabetes is one of the popular chronic diseases, the health experts phrased it, “Disease of the future”, as the diabetes cases are rising and have been growing yearly [8]. According to the Immune Deficiency Foundation (IDF), 537 million people are suffering with diabetes today, with almost 1 out of 2 cases being undiagnosed. By 2045, this figure is expected to soar to a staggering 783 million. The symptoms, physiological body parameters, complications, and other circumstances are different for each and every patient, the physician will suggest different prescription, diagnosis, and course of action to each patient, which can lead to discrepancies and inconsistency in the quality-of-care services [9] [10]. Various researches have figured out the risk factors related to diabetes that can help in diagnosis, treatment and to monitor these inconsistencies in the diagnostic process[11] [12]. Optical character recognition (OCR) can read both handwritten and printed text [13]. The potential of the input documents will directly affect the OCR performance [11]. OCR also allows us to convert any documents into editable and searchable data [14]. Ontology is a principal tool for the structuring and representation of knowledge [12], [15]. Ontologies can reflect the meaning of a scientific domain and facilitates the sharing of domain knowledge between people and software. They can compile different patients related data likewise ailments, diagnoses, treatments, medications, and enable the development of a treatment plan as per the needs of the patient and also perform the modification of clinical decision support system [12]. An ontology-based approach has theoretical and practical benefits in developing automated algorithms to identify people with chronic diseases (like diabetes, hypertension, obesity etc.). They are also helpful in creating therapeutic standards and care regimens [16]. Cyber Physical Systems are also very popular these in handling the challenges of real time healthcare data. With the help of these systems various data management issues can be solved to provide the good quality healthcare services [17].
