Original Articles

Factors associated with macrosomia among singleton live-birth: A comparison between logistic regression, random forest and artificial neural network methods


Abstract


Background: Macrosomia with adverse outcomes for mother and infant is affected by several risk factors. Classification methods including logistic regression, random forest and artificial neural network are used to determine high risk groups for macrosomia.

 

Methods: This cross-sectional study was conducted on 4342 pregnant women who gave singleton live-birth in Tehran, Iran from 6-21 July 2015. The performed methods were compared using tools such as sensitivity, specificity and accuracy. McNemar's test checked differences in proportion among methods. To assess the association between the observed and predicted values, Kappa statistic was calculated.

 

Results: Mother’s BMI, SES, mother’s education, parity, mother’s age, gestational age and mother’s occupation are the most important variables affecting macrosomia identified by RF method with the highest accuracy 0.89. The association of RF predictions and observed values using Ø coefficient, contingency coefficient, Kendall tau-b and kappa were 0.43, 0.39, 0.43 and 0.31, respectively.

 

Conclusion: Based on our findings, random forest had the best performance to classify macrosomia comparing to artificial neural network and logistic regression and may be used as an appropriate method in such data.


Full Text:

PDF


NBN: http://nbn.depositolegale.it/urn%3Anbn%3Ait%3Aprex-19076

DOI: http://dx.doi.org/10.2427/11985

References



Article Metrics

Metrics Loading ...

Metrics powered by PLOS ALM

Refbacks

  • There are currently no refbacks.


Copyright (c) 2013 Epidemiology, Biostatistics and Public Health

Creative Commons License
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

EBPH Epidemiology, Biostatistics and Public Health | ISSN 2282-0930

Creative Commons License
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.