Application of computational technique in design of classifier for early detection of gestational diabetes mellitus

Abstract

Gestational Diabetes Mellitus (GDM) is defined as any degree of glucose intolerance with onset or first recognition during pregnancy. In view of maternal morbidity and mortality as well as fetal complications, early diagnosis is an utmost necessity in the present scenario. In developing country like India, early detection and prevention will be more cost effective. Oral Glucose Tolerance Test (OGTT) is the crucial method for diagnosing GDM done usually between 24th and 28th week of pregnancy. The proposed work focuses on early detection of GDM without a visit to the hospital for women who are pregnant for the second time onwards (multigravida patients). A decision support system using Multilayer Neural Network which learns to classify GDM and non GDM patients using Back 3328 Priya Shirley Muller et al. Propagation learning algorithm is developed. The classifier proves to be an efficient model for diagnosis of GDM without the conventional method of blood test by providing newly designed parameters as inputs to the network.

A Review of Epidemic Forecasting Using Artificial Neural Networks

“…20 ANN is an important tool for data mining of medical records for classification and prediction purposes. 22 In a large number of previous studies, neural network was used for classifying such diseases as dengue fever, [23][24][25] chest or heart diseases, 26,27 West Nile virus diseases, 28 tuberculosis, 29,30 gestational diabetes mellitus, 31 swine flu, 32 and pancreatic cancer. 33 These studies had helped in diagnosis and case management of epidemic victims.…”

Effects of pre-pregnancy maternal body mass index on gestational diabetes mellitus

“…Fourth to eighth variables deal with previous pregnancy information such as presence of GDM, birth of a baby which weighed more than 3.8Kg, death of a baby before 20 weeks, birth of a baby with defects in spinal cord, heart or brain, death of a baby after 20 weeks respectively. The last two reveal information on history of urinary, skin or vaginal infections and presence of polycystic ovary syndrome [7]. Eight of the ten variables used are binary variables, where 0 indicates nonoccurrence and 1 indicates occurrence.…”

Optimization of Backpropagation for Early Detection of Diabetes Mellitu

“…Studies concerning this matter have been carried out in several methods, such as combining Regression Tree and Random Forest (RF) [4], Fuzzy Hierarchical Model [5], Genetic Programming [6], Support Vector Machines (SVM), Naïve Bayes [7] and artificial neural network [8], [9]. Input data to be in [10], [11], face area [12], [13] and magnetic resonance imaging of the brain [14].…”

“…Voice data may also be included as a data input based on several parameters that consist of absolute jitter, shimmer, amplitude perturbation quotient, noise-toharmonic ratio, smoothed amplitude perturbation quotient and relative average perturbation [15]. Artificial neural network is an excellent method to diagnose disease [8], [9], [16][17][18][19][20]. Jayalaksmi and Sansthakumaran point out that artificial neural network may be implanted in diagnosing diabetes mellitus and classifying the early detection of gestational diabetes mellitus [8].…”

“…The active parameters involve the number of pregnant times, plasma glucose concentration, blood pressure, triceps skin fold thickness, insulin serum, body mass index, diabetic pedigree function and age. In another study, backpropagation was employed to classify the early detection of gestational diabetes mellitus [9]. This study observed 110 data and promoted 10 parameter inputs: family history of diabetes, pre-pregnancy body mass index, history of gestational diabetes, delivery of a large infant, history of miscarriage, abnormal baby in previous pregnancy, history of stillbirth, infections, and history of polycystic ovary syndrome.…”