Abstract:
In this research we explored the potential of the multivariate methods, CCA and PLS regression in studying the relationship between two sets of variables in public health research. This research will help the public health researchers in choosing the appropriate statistical methods if they want to study multiple outcomes and multiple exposures simultaneously. Additionally, we intended to identify a surrogative measure of low birth weight (LBW) using ROC curve and 4 machine learning algorithms, decision tree, random forest, support vector machine and neural network. Both the canonical correlation analysis and the PLS regression analysis has several advantages over the univariate methods. However, CCA is just an exploratory method very similar to Pearson’s correlation. Although one variable set is often considered as predictor and the other as criterion, it does not imply causal relationship between the set of exposures and the set of outcomes. On the other hand, PLS regression can help in establishing causal relationship between exposures and the outcomes. So, the choice of CCA or PLS regression depends on the objective of the study. In addition, this study will provide a realistic predictive model of LBW for Rural Bangladesh. The LBW can be predicted with 4 other simpler-to-measure anthropometries, length and head, chest and arm circumferences without measuring their weights, at a greater accuracy with the advent of the information technology.
