Article Preview
Top1. Introduction
The most important thing in the medical field is proper and accurate diagnosis as it is challenging even for a well-qualified doctor having ample experience in handling thousands of comparable cases. In many cases, the existing symptom of an infectious disease is an indicator for other problems as well. Moreover, tracing the root cause becomes strenuous due to the co-occurrence of other complex syndromes. For example, even the diagnosis of a patient suffering from a common fever, requires significant scrutiny, as most of the patients have a history of diabetes, hypertension and other related co-morbidity. In light of this, diagnosis involves exploring a whole gamut of links in the chain. Consequently, the possible state space of diagnosis landscape becomes significantly large, which makes it extremely difficult to deal with, especially when resource availability is scarce. On the other hand, to lessen this possibility, the concerned specialist sticks to a particular decision sequence for a given illness trajectory of the patient. However, it increases the risk of loss of life due to unidentified illness and also non-correctable errors in a timely manner.
Complexity of the diagnostic process is increased due to the imprecision and vagueness involved with the underlying data as patients are not able to express properly their health conditions, thus not allowing healthcare practitioners to interpret the information correctly. This leads to the high levels of imprecision involved in the measuring instruments, ambiguous nature of the symptoms, disease prognosis due to environmental hazards, which is true in case of dengue fever. Early diagnosis and immediate medical attention assist in controlling the epidemic and also restraining the mortality and morbidity rate. Rather than restricting the process of diagnosis to the medical practitioners’ judgement only, intelligent decision support system assisting doctors in strengthening their decision-making ability (Musen, Shahar, & Shortliffe, 2001), building confidence in less experienced professionals above the human judgement error, can be explored as specialist doctors hard to avail around the clock. In addition to assisting doctors, the system must handle the complex and diverse symptoms associated with the tropic disease. Hence soft computing techniques are quite suitable for these kind of problems (Oguntimilehin, Adetunmbi, & Abiola, 2013; Samuel & Omisore, 2013; Gambhir, Malika, & Kumar, 2016). Several research employ multiple soft computing intelligent techniques for investigating tropical diseases like malaria (Djam et al., 2011) and dengue (Adawiyah, 2017; Afan, Yen, & Christian, 2014; Gambhir, Malika, & Kumar, 2017).
1.1 Motivation for the Study
Dengue fever is the most common mosquito-borne viral illness in the world, endemic in more than 100 countries, with 2.5 billion people at risk for transmission. Most prevalent in tropical and subtropical areas where there is limited public health infrastructure for mosquito control in areas such as: Africa, Asia, Mexico, and central and south America. The widespread outbreak of dengue, usually in under-developed countries, mainly in monsoon season, poses a challenge for the administration, in terms of treatment with available resources as it is not feasible to scale-up the required healthcare resources in a short notice. Consultation with a physician or a specialist if required is not available in a timely manner in remote areas or even in urban places due to high numbers of patients waiting for consultation. In addition to this, general hospital having proper nursing care, laboratory, and intensive care unit (ICU) facility cannot be multiplied overnight. Besides laboratory tests, dengue serology testing by Elisa kits, is found to be too expensive. In the case of dengue, most importantly the fluid level in the body is maintained, platelet level is constantly monitored and clinical care is provided since no major medication, and surgery is performed in treatment except administering general intravenous antibiotics. In the case of scarcity of doctors, suitable laboratory tests, nurses and other paramedical staff come forward to take the responsibility in the shoulder. In this case, a clinical diagnosis by artificial intelligence techniques comes to the rescue of nursing staff and even less-experienced physicians and house surgeon staff during making an initial assessment about the severity of occurrence of disease.