小湿妹视频

Early Diagnosis of Diabetes Mellitus in Oman using AI- based Predictive Algorithm

Diabetes mellitus is observed as the fastest growing disease in the world. By 2050, the world’s diabetes patients are expected to reach to over 700 million, which means one in 20 adults will be suffering from diabetes according to “World Population Ageing report” published by the United Nation in 2015. The Oman National Center of Data and Statistic reports that diabetes mellitus is ranked 4th as the cause of death in Oman in 2019. Therefore, early diagnosis and treatment will minimize mortality due to diabetes.

The main aim of the research is to develop a predictive model for early diagnosis of diabetes mellitus type II among Omani nationals using AI-based techniques such as machine learning (ML) support vector machine (SVM), convolutional neural networks (CNN), deep learning (DL) and combined ML and DL. The early diagnosis of the disease is vital to provide early treatment intervention to control disease progression and minimize premature death.

This research provides a common understanding of diabetes mellitus classification using artificial intelligence, machine learning and deep learning. The research began by reviewing all relevant studies and explored the accuracy in diabetes prediction. At first, the published studies were analyzed in detail and classified according to their methodologies. The comprehensive and detailed review of the diagnosis of diabetes by machine learning algorithms as well as the combined models have been compiled into a comprehensive report. The literature review also includes the accumulation and creation of classification and prediction techniques. The published prediction models use different types of machine learning algorithms such as classification or association algorithms; Decision Trees, Support Vector Machine (SVM), and Linear Regression. They were the most common algorithms used until July 2020. Deep Learning (DL) has been introduced as an improvement to ANN. Recent studies that have used DL produced remarkable results. The accuracy rate produced by these methods varied. This has encouraged us to attempt to improve the accuracy by either building models with classifiers that haven’t been used or combine different classifiers. The majority of the studies in the field of the diabetes prediction used the public Pima Indian Dataset obtained from the UCI repository. We are working closely with clinical collaborators in Oman to collect the available Diabetic Clinic data in different regions for age groups above the age of 40 as the national screening is routinely done for this age group in Oman. This data will be preprocessed and used to develop the AI predictive model.