Figure 1.

Classifiers exhibit greater accuracy increase from baseline when stratified by sex. (a) Analyses on the overall dataset yielded up to 6% greater accuracy than the baseline accuracy (56.5%) using the support vector machine (SVM) classifier and the joint mutual information (JMI) feature selection. (b) Prediction of male seasonality yielded up to 10% greater accuracy than the baseline accuracy (58%) using the RF classifier and correlation-based feature selection. (c) Prediction of female seasonality yielded up to 7% greater accuracy than the baseline accuracy (62.2%) using the LR and SVM classifiers and mRMR feature selection.