Title :
A modeling approach to identify factors associated with infant mortality in Russia
Author :
Soshnikov, Sergey ; Lee, Chi-Kwan ; Vladimirov, Sergey
Author_Institution :
Dept. of Med. & Social Problems, CPHRI, Moscow, Russia
Abstract :
The infant mortality has been declining in the recent decades globally. However, it continues to be an international concern. The issues are not only on what types of diseases or incidents leading to the infant death for individual cases, but also on what factors; social, economic, environment, education and other; are associated with infant mortality in the society. In this study, we collect infant mortality data along with over 100 variables using region as the sampling unit in Russia to investigate potential factors that may have significant association with infant mortality in Russia. Our purpose has two folds. First is to address the issues of data quality and the selection of modeling methodology for observational studies in general. Second is to identify possible important factors associated with infant mortality in Russia. We conclude with a discussion of pitfalls and cautions when applying predictive modeling techniques in observational studies.
Keywords :
data analysis; diseases; sampling methods; socio-economic effects; Russia; data quality; diseases; economic factors; educational factors; environment factors; infant death; infant mortality data; predictive modeling techniques; social factors; Biological system modeling; Buildings; Data models; Input variables; Predictive models; Sociology; Statistics; Data Quality; Decision Tree; Ensemble; Gradient Boosting; LARS; LASSO; Neural Network; Partial Least Squares;
Conference_Titel :
Computer and Information Science (ICIS), 2013 IEEE/ACIS 12th International Conference on
Conference_Location :
Niigata
DOI :
10.1109/ICIS.2013.6607838