Article Preview
TopIntroduction
Diabetes is a common health problem and is often described by professionals as diabetes mellitus (DM). It comprises a group of metabolic disorders that manifest in hyperglycemia (high blood sugar) in the body of the sufferer either due to insufficient insulin production, insulin intolerance, or both. It has been suggested that it is best to diagnose DM at the early stages. The World Health Organization (WHO) report of November 14, 2016, estimated that about 422 million people had diabetes, while 1.6 million deaths are associated with DM. Based on this report, it is easy to predict the severity of diabetes in patients worldwide.
About 8.5% of adults over 18 years old had diabetes in 2014, and in 2012, hyperglycemia was the primary cause of death (about 2.2 million deaths were reported). The onset of DM comes with various levels of damage to different body parts, such as the eyes, kidneys, nerves, and heart. According to the Williams Textbook of Endocrinology, more than 382 million people had DM globally in 2013, and many of them died due to DM-related issues in both poor and wealthy countries of the world.
The US Centers for Disease Control and Prevention (CDC) predicted a 23% increase in type-II DM cases for nine years straight (2001–2009). This disease has become a source of worry to leaders of many countries and organizations, especially relating to its prevalence and prevention. Diabetes is classified into two types: type I and type II. Type I diabetes is commonly described as insulin-dependent diabetes because of the inability of the human body to produce enough insulin. This class accounts for about 10% of all diabetes cases. Type II diabetes, as per the Canadian Diabetes Association, is expected to increase from 2.5 million cases to 3.7 million cases between 2010 and 2020. Hence, diabetes is a health condition that requires early diagnosis and prevention to avoid its life-threatening consequences.
In the last few decades, many health researchers have utilized machine-learning algorithms for the prediction and classification of many diseases. Fast and accurate prediction allows for the development of early and effective treatments. For instance, Abdar et al. (2017) used two well-known data-mining-based algorithms for the detection of liver disease. Their work showed that the algorithms Boosted C5.0 and CHAID are capable of producing new rules for detecting liver disease risk factors.
Moreover, the availability of data and the rise of computational capabilities have encouraged scientists to analyze clinical data to find answers to these pressing problems. This is done using data-mining techniques to discover useful information from large health datasets. Advancements in information technology have made data mining a useful tool in diabetes studies as it has led to the improvement of healthcare delivery and increased decision-making support for better disease supervision. Aljumah et al. (2013) concentrated on predictive analyses of diabetic treatment using regression-based data-mining techniques. They employed Oracle Data Miner (ODM) software as a mining tool for predicting the mode of treating diabetes. In this study, the target variable’s identification was based on their percentage. Patients’ treatment processes were also considered. Patients were grouped into categories (old or young) based on their age before predicting their treatment. This study observed high predictive percentages for both the young and old control groups. The treatment predictive percentage was calculated using support vector machines (SVM).
Currently, there is no single technique that offers the highest accuracy in predicting all diseases since the excellent performance of a given classifier on one disease dataset can be outdone in another disease dataset. In a study by Bashir et al. (2016), a new hybridization of different classifiers was proposed for the prediction and classification of diabetes. This hybridization approach was proposed to overcome the issues associated with each of the classifiers. In the next section, related work is presented and discussed. Then we introduce the methodology, results, and discussion, and we offer some closing thoughts in the conclusion.