Predicting Diabetes Mellitus With Machine Learning Techniques Using Multi-Criteria Decision Making

Predicting Diabetes Mellitus With Machine Learning Techniques Using Multi-Criteria Decision Making

Abhinav Juneja, Sapna Juneja, Sehajpreet Kaur, Vivek Kumar
Copyright: © 2021 |Pages: 15
DOI: 10.4018/IJIRR.2021040103
OnDemand:
(Individual Articles)
Available
$37.50
No Current Special Offers
TOTAL SAVINGS: $37.50

Abstract

Diabetes has become one of the common health issues in people of all age groups. The disease is responsible for many difficulties in lifestyle and is represented by imbalance in hyperglycemia. If kept untreated, diabetes can raise the chance of heart attack, diabetic nephropathy, and other disorders. Early diagnosis of diabetes helps to maintain a healthy lifestyle. Machine learning is a capability of machine to learn from past pattern and occurrences and converge with experience to optimise and give decision. In the current research, the authors have employed machine learning techniques and used multi-criteria decision-making approach in Pima Indian diabetes dataset. To classify the patients, they examined several different supervised and unsupervised predictive models. After detailed analysis, it has been observed that the supervised learning algorithms outweigh the unsupervised algorithms due to the output class being a nominal classified domain.
Article Preview
Top

Introduction

Diabetes is the fastest growing disease amid the individuals even among the youngsters. Several studies have been done to investigate the extent of diabetes in India and their main focus was on avoidance of this disease (Kaveeshwar & Cornwall, 2014). Diabetes mellitus (DM) or simply diabetes is a group of inter related metabolic diseases in which a person has high blood sugar (Abhari et al., 2019). This high blood sugar exhibit the symptoms of recurrent urination, dehydration and sudden increased hunger. If kept untreated, diabetes can cause many problems. The other related complications that may trigger with Diabetes include diabetic ketoacidosis and non-ketotic hyperosmolar coma (Lam et al., 2016). Some of the most severe and serious long-term complications pertaining to the after effects of include heart disease, kidney failure, and improper vision leading to even total loss. (Saran et al., 2015; Jain, A. and Bhatnagar, 2016)

Diabetes due to its symptoms and the extent of spread and origin has been classified into various types (Contreras & Vehi, 2018). Three of the most dominant diabetes forms elaborated in the work are as follows:

  • 1.

    Type 1 Diabetes: Type 1 diabetes may also be referred to as insulin-dependent, teenaged or childhood diabetes which is characterised by scarcity in of production of insulin in the body and further necessitates management and control of insulin on a day to day basis (Kavakiotis et al., 2017). The origin and stimulus for type 1 diabetes is unexplored and it is not curable with current diagnosis and treatment (Kitabchi et al., 2009). Typical Symptoms of type 1 diabetes involves excessive secretion of urine also known as polyuria, dehydration or polydipsia, continuous feeling of hunger, loss in the body weight, change in vision, and drowsiness. These symptoms can arise instantly without much notice.

  • 2.

    Type 2 Diabetes: Type 2 diabetes also known as non-insulin-dependent, or adult-onset diabetes, occurs because of the human body’s unproductive utilization of insulin (Wu et al., 2014). Type 2 diabetes relates to most of the people suffering from diabetes across the globe, and is mainly the result of excessive weight of the body and lack of physical activity. Symptoms are somehow indistinguishable from those of type 1 diabetes. As a consequence, the disease may be identified several years after inception, once symptoms or complications come to light. Few years back, this type of diabetes was recognised only in adults but it is now observed in the children also.

  • 3.

    Gestational Diabetes: Gestational diabetes is state of hyperglycaemia with blood glucose values exceeding normal, but less than normal in patients if diagnosed during pregnancy. Females suffering from gestational diabetes are prone to higher risk of complications during pregnancy and at the time of child birth. They and their children are also at increased risk of type 2 diabetes in the future. Gestational diabetes is diagnosed through prenatal screening, instead of reported symptoms.

There is a continuous quest on why the blood glucose level becomes high? Due to low levels of insulin in body, blood glucose level rise steadily. Insulin is a hormone that is generated by pancreas, and it permits our body to utilize sugar from carbohydrates, from the food that anyone consumes for energy or to reserve glucose for later use. Insulin supports the body by maintaining the blood sugar level and controls it from bring excessively high or extremely low. The human body cells need sugar as source of energy, while sugar cannot pass into maximum cells of the body directly. After the consumption of food, blood sugar level rises and the pancreas cells known as beta cells, activate to discharge insulin into the blood vessels. Insulin then signals cells to consume sugar from these blood vessels. Insulin is often represented as a “key”, which opens up the cell to permit sugar to get into the cell and be utilized for obtaining energy (Saedi et al., 2016). If there is excess of sugar then required, insulin helps in storing that in the liver, which may be further used as and when required by body during workout or any physical activity. With rising blood sugar levels, pancreas generates additional insulin. If the body of patient is not capable of producing sufficient amount of insulin, the patient has the possibility of developing high blood sugar or hyperglycemia in medical terms, which can lead to long-term problems if the blood sugar level continues to rise for a long duration.

Complete Article List

Search this Journal:
Reset
Volume 14: 1 Issue (2024)
Volume 13: 1 Issue (2023)
Volume 12: 4 Issues (2022): 3 Released, 1 Forthcoming
Volume 11: 4 Issues (2021)
Volume 10: 4 Issues (2020)
Volume 9: 4 Issues (2019)
Volume 8: 4 Issues (2018)
Volume 7: 4 Issues (2017)
Volume 6: 4 Issues (2016)
Volume 5: 4 Issues (2015)
Volume 4: 4 Issues (2014)
Volume 3: 4 Issues (2013)
Volume 2: 4 Issues (2012)
Volume 1: 4 Issues (2011)
View Complete Journal Contents Listing