Mario W. L. Moreira

Evolutionary radial basis function network for gestational diabetes data analytics

Abstract The development of smart decision support systems (DSSs) that seek to simulate human behavioral aspects is a major challenge for computational intelligence (CI). Artificial neural network (ANN) approaches have the ability to solve complex decision-making problems that involve uncertainty and a large amount of information in a fast and reliable way. The application of this evolutionary CI technique to analyze a large amount of data is an important strategy to solve several problems in healthcare management. This paper proposes the modeling, performance evaluation, and comparison analysis of an ANN technique known as the radial basis function network (RBFNetwork) to identify possible cases of gestational diabetes that can lead to multiple risks for both the pregnant women and the fetus. This method achieved promising results with a precision of 0.785, F -measure of 0.786, ROC area of 0.839, and Kappa statistic of 0.5092. These indicators show that this ANN-based approach is an excellent predictor for gestational diabetes mellitus. This research provides a comprehensive decision-making model capable of improving the care provided to women who are at a risk of developing gestational diabetes, which is the most common metabolic problem in gestation with a prevalence of 3–18%. Thus, this work can contribute to the reduction of maternal and fetal mortality and morbidity rates.

A preeclampsia diagnosis approach using Bayesian networks

Hypertension is the main cause of maternal death. Preeclampsia can affect pregnant women before or during pregnancy. Identification of patients with higher risk for preeclampsia allows some precautions that are taken to prevent its severe disease and subsequent complications. In medicine, there are different situations that deal with a large range of information, which needs a thorough assessment to be able to help experts in the decision-making process. Smart decision support systems allow grouping all existing information and finding pertinent information from it. Bayesian networks offer models that allow the information capture and handle situations of uncertainty. This paper proposes the construction of a system to support intelligent decision applied to the diagnosis of preeclampsia using Bayesian networks to help experts in the pregnants care. The processes of qualitative and quantitative modeling to the construction of a network are also presented. The main contribution of this work includes the presentation of a Bayesian network built to help decision makers in moments of uncertainty in care of pregnant women.

An inference mechanism using Bayes-based classifiers in pregnancy care

Significant advances on smart decision support systems (DSSs) development have influenced important results on pregnancy care. Nevertheless, even considering the efforts to reduce the number of women deaths due to problems related to pregnancy, this decrease presented less impact than other areas of human development. Hypertensive disorders in pregnancy, particularly pre-eclampsia and eclampsia, account for significant proportion of perinatal morbidity and maternal mortality. In this context, this paper proposes an inference model that uses data mining (DM) techniques capable for operating in a data set to extract patterns and assist in knowledge discovery. Identifying hypertensive crises that complicate pregnancy, it can impact in a meaningful reduction the incidence of sequelae and death of pregnant women. Comparison between two Bayesian classifiers is performed in this work to better classify the hypertensive disorders severity. Results showed that Naïve Bayes classifier had an excellent performance, presenting better precision and F-measure, compared to the other experimented classifiers. Even finding a good performance to predict hypertensive disorders, other Bayesian methods need to be evaluated, as well as other DM techniques such as those based on artificial intelligence (AI) and tree-based methods.

Nature-Inspired Algorithm for Training Multilayer Perceptron Networks in e-health Environments for High-Risk Pregnancy Care

Nature presents an infinite source of inspiration for computational models and paradigms, in particular for researchers associated with the area known as natural computing. The simultaneous optimization of the architectures and weights of artificial neural networks (ANNs) through biologically inspired algorithms is an interesting approach for obtaining efficient networks with relatively good generalization capabilities. This methodology constitutes a concordance between a low structural complexity model and low training error rates. Currently, complexity and high error rates are the leading issues faced in the development of clinical decision support systems (CDSSs) for pregnancy care. Hence, in this paper the use of a biologically inspired technique, known as particle swarm optimization (PSO), is proposed for reducing the computational cost of the ANN-based method referred to as the multilayer perceptron (MLP), without reducing its precision rate. The results show that the PSO algorithm is able to improve computational model performance, showing lower validation error rates than the conventional approach. This technique can select the best parameters and provide an efficient solution for training the MLP algorithm. The proposed nature-inspired algorithm and its parameter adjustment method improve the performance and precision of CDSSs. This technique can be applied in electronic health (e-health) systems as a useful tool for handling uncertainty in the decision-making process related to high-risk pregnancy. The proposed method outperformed, on average, other approaches by 26.4% in terms of precision and 14.9% in terms of the true positive ratio (TPR), and showed a reduction of 35.4% in the false positive ratio (FPR). Furthermore, this method was superior to the MLP algorithm in terms of precision and area under the receiver operating characteristic curve by 2.3 and 10.2%, respectively, when applied to the delivery outcome for pregnant women.

Predicting hypertensive disorders in high-risk pregnancy using the random forest approach

The incidence of hypertension associated with pregnancy contributes significantly to increase maternal and fetal deaths during pregnancy and childbirth. Due to its high incidence rate and several complications, the study of this disorder has been subject of numerous investigations in an attempt to determine its prevention and improve the treatment conduction. In this context, this paper uses a data mining (DM) technique, named random forest (RF), applied to health care to early identification of these disorders. It also presents the modeling, performance assessment, and comparison with other DM methods to evaluate the performance of the proposed model. Results showed that the RF classifier had a regular performance, presenting the best values for true positive Rate (TP Rate) and recall in the prediction of preeclampsia superimposed on chronic hypertension compared to the other experimented classifiers. Even finding a good performance to predict hypertensive disorders, other tree-based methods need to be evaluated, as well as other DM techniques. Discovering reliable information of pregnant women suffering from the hypertensive disease is an important path to reduce the high rate of deaths, mainly, in developing countries where 99% of these deaths occur.

Smart mobile system for pregnancy care using body sensors

Hypertensive disorders are the most common problems during pregnancy. They cause about 10% of maternal deaths. The world mortality rate has decreased but many women are still dying every day from pregnancy complications. Various technic resources are being used in an integrated manner in order to minimize even more the death of both mothers and babies. Mobile devices with Internet access have a great potential to expand actions of health professionals. These devices facilitate care with people that are living in remote areas, assisting in patient monitoring. Information exchange anywhere and anytime between experts and patients could be an important way to improve the pregnancy monitoring. This paper presents a mobile monitoring solution using body sensors to identify worsens in the health status of pregnant women suffering hypertensive disorders. This mobile application uses Naïve Bayes classifier to better identify hypertension severity helping experts in decision-making process. Results show that the proposed mobile system is promising for monitoring blood pressure disorders in pregnancy.

Postpartum depression prediction through pregnancy data analysis for emotion-aware smart systems

Abstract Emotion-aware computing represents an evolution in machine learning enabling systems and devices process to interpret emotional data to recognize human behavior changes. As emotion-aware smart systems evolve, there is an enormous potential for increasing the use of specialized devices that can anticipate life-threatening conditions facilitating an early response model for health complications. At the same time, applications developed for diagnostic and therapy services can support conditions recognition (as depression, for instance). Hence, this paper proposes an improved algorithm for emotion-aware smart systems, capable for predicting the risk of postpartum depression in women suffering from hypertensive disorders during pregnancy through biomedical and sociodemographic data analysis. Results show that ensemble classifiers represent a leading solution concerning predicting psychological disorders related to pregnancy. Merging novel technologies based on IoT, cloud computing, and big data analytics represent a considerable advance in monitoring complex diseases for emotion-aware computing, such as postpartum depression.

Semantic interoperability and pattern classification for a service-oriented architecture in pregnancy care

Abstract Semantic interoperability represents one of the main challenges in health information systems. The development of novel interoperability models should promote the integration of heterogeneous information in the acquisition and semantic analysis of complex data patterns, which are typically used in clinical information. The purpose of this study is to develop a knowledge-based decision support system that uses ontologies for integrating data related to hypertensive disorders in pregnancy. This model allows, when dealing with new cases, inferring from a knowledge base and predicting high-risk situations that could lead to serious problems during gestation in both pregnant women and fetuses. Results demonstrate that the use of ontologies to address semantically acquired patterns from different electronic health records has the potential to significantly influence a service-oriented architecture implementation for clinical decision support systems.

Multilayer Perceptron Application for Diabetes Mellitus Prediction in Pregnancy Care

The human intelligence modeling by brain components simulation, such as neurons and their connections, is part of leading smart decision computing paradigms. In Health, artificial neural networks (ANN) have the capacity to adapt to uncertainty situations and learn even with inaccurate data. This paper presents the modeling and performance evaluation of an ANN-based technique, named multilayer perceptron (MLP), for gestational diabetes mellitus (GDM) prediction that is responsible for several severe complications and affects 3 to 7% of pregnancies worldwide. Results show that this approach reached a precision of 0.74, Recall 0.741, F-measure 0.741, and ROC area 0.779. These indicators show that this method is an excellent predictor of this disease. This contribution offers a computational intelligence (CI) tool capable of identifying risk cases during pregnancy and, thus, reduce possible sequels for both pregnant woman and fetus.