A Predictive Analytic Model for Maternal Morbidity

A Predictive Analytic Model for Maternal Morbidity

Edgardo Palza (École de technologie supérieure, Canada), Jorge Sanchez (Universidad Peruana Unión, Peru), Guillermo Mamani (Universidad Peruana Unión, Peru), Percy Pacora (Hospital Nacional Docente Madre Niño “San Bartolomé”, Peru), Alain Abran (École de Technologie Supérieure, Canada) and Jane Moon (University of Melbourne, Australia)
DOI: 10.4018/978-1-4666-9432-3.ch005
OnDemand PDF Download:
No Current Special Offers


This chapter presents a predictive analytic model for preventing neonatal morbidity through the analysis of patterns of risky behavior regarding morbidity in newborns. The chapter presents the design and implementation of a forecasting model of Neonatal morbidity. The model developed is based on artificial intelligence using Bayesian Networks, Influence Diagrams and principles of traditional statistics. The model research is based on a repository of 10,000 medical records at a hospital in Peru. The model aims to identify the factors that are causes of morbidity in newborns, is based on data mining techniques and developed using the CRISP-DM methodology.
Chapter Preview

The Empirical Context And The Development Phases For The Predictive Analytic Model

Empirical Context

The HNDSB has a Perinatal Information System (PIS) which records the basic history of the mother and child, including all tests and controls that have been undertaken from the beginning of pregnancy until the time of delivery. PIS was one of the most important sources of data for the Analytic Model, while others were social reports. The population available in the PIS database consisted of about 65,000 medical records of pregnant patients collected between 2000 and 2010.

Complete Chapter List

Search this Book: