Infant mortality is an important health measure in a population as a crude indicator of the poverty and socioeconomic level. It also shows the availability and quality of health services and medical technology in a specific region. Although improvements have been observed in the last decades, the implementation of actions to reduce infant mortality is still a concern in many countries. To address such an important problem, this paper proposes a new support decision approach to classify newborns according to their neonatal mortality risk. Using features related to mother, newborn, and socio-demographic, we model the problem using a data-driven classification model able to provide the probability of a newborn dying until days of life. More than a theoretical study, decision support tools as the one proposed here is relevant in countries in development as Brazil, because it aims at identifying risky neonates that may die to raise the attention of medical practitioners so that they can work harder to reduce the overall neonatal mortality. Overcoming an AUC of 96%, the proposed method is able to provide not just the probability of death risk but also an explicable interpretation of most important features for model decision, which is paramount in public health applications. Furthermore, we provide an extensive analysis across different rounds of experiments, including an analysis of pre and post partum features influence over data-driven model. Finally, different from previously conducted studies which rely on databases with less than 100,000 samples, our model takes advantage from a new proposed database, constructed using more than 1,400,000 samples comprising births and deaths extracted from public records in São Paulo-Brazil from 2012 to 2018.

Artificial Neural Networks (ANN) have been frequently applied to reduce risks and maximize the net returns in different types of algorithm trading. Using a real dataset, and aiming to support the Market Making process in High-Frequency Trading, this work investigates the use of a multilayer perceptron (MLP) to predict positive oscillations in short time periods (5, 10 or 15 minutes). The statistical analysis of our results showed that a neural network is more effective in short-term oscillations (5 minutes) when compared with the results obtained in longer periods (10 or 15 minutes). The result is important because it allows to insert a higher quantity of limit orders once they will be placed more frequently, which increases the market liquidity. It contextualizes a new contribution in the High-Frequency Trading field, where this work proposes a new trigger to start a market making process.

No abstract