Abstract
Diabetes mellitus, a metabolic disease that features high glucose levels in the body with the inability of the body to secret enough insulin to breakdown glucose, or such a body is resistant to the effects of insulin. Nigeria and other nations of the world have become aware of the inherent threats to life of gestational diabetes in mothers with or without previous cases and its tendencies to metamorphose into Type-II. Our study presents a comparative study of classification models using both the supervised (K-nearest neighborhood and Quadratic Discriminant Analysis) and unsupervised (Profile Hidden Markov Model and Memetic algorithm) methods – which aims at early detection as well as improve early diagnosis via data-mining tools. Adopted dataset is split into: training (in some cases, retraining) and testing to aid model validation. Results show that age, obesity and family ties to the second degree, environmental conditions of inhabitance are critical factors that can increase likelihood. Gestational diabetes in mothers with or without previous cases were confirmed if: (a) history of babies weighing > 4.5kg at birth, (b) insulin resistance with polycystic ovary syndrome, and (c) abnormal tolerance to insulin. Also, PHMM outperforms Memetic algorithm in some cases; while memetic algorithm outperforms PHMM in some cases.
Highlights
Diabetes mellitus has become a general chronic disease that affects about 6% of the global population – so that its avoidance and early detection for effective treatment has become imperative and undoubtedly a critical task for health and economic issue in 21st century (Khashei et al, 2012)
Gestational Diabetes Mellitus (GDM) is defined as disorder of glucose tolerance occurring first in pregnancy in mothers – whereas, some experts have viewed and believe GDM to be of same entity with Type-II – wherein the former constitutes the early signs and manifestation of the latter
If and when there is a large difference between covariance matrices, and a small difference between sample proportions, the optimal value k is determined by N2/8 (Enas and Choi, 1986). This model presents several merits (Berrueta et al, 2007) in that: (a) its mathematical simplicity does not prevent it from achieving classification results as good as other more complex pattern recognition techniques, (b) it is free from statistical assumptions, (c) its effectiveness does not depend on the space distribution of the classes, and (d) when the boundaries between classes are not hyper‐linear or hyper‐ conic, K-nearest neighbour performs better than Linear Discriminant Analysis (LDA)
Summary
Diabetes mellitus has become a general chronic disease that affects about 6% of the global population – so that its avoidance and early detection for effective treatment has become imperative and undoubtedly a critical task for health and economic issue in 21st century (Khashei et al, 2012). Diabetes is a metabolic disease that is characterized by the presence of hyperglycemia or high blood glucose. Ojugo et al (2015) Various models have been used for its early detection and identification to include: (a) supervised classification in which its input variables for the diagnosis are known), and (b) unsupervised classification in which the variables used for diagnosis and classification are unknown). In both instance, a critical feat in selecting the appropriate classification model to use is, its accuracy and precision ability in classifying the task at hand
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
