Artificial neural network prediction of Indonesian mothers' emotional well-being: The role of social support, quality of life, and sharenting behavior.

Abstract

Within the framework of the stress-buffering model and ecological systems theory, the increasing integration of digital engagement into motherhood underscores the need to understand how psychosocial and technological factors relate to maternal mental health. This study examined the associations of demographic factors, social support, quality of life (QoL), and sharenting behavior with Indonesian mothers' emotional well-being (EWB). A cross-sectional survey was conducted with 300 Indonesian mothers aged 21-55 years who had children aged 0-8 years. Pearson correlations were used to examine bivariate associations, and an artificial neural network (ANN) was used to classify participants into below-median and above-median EWB groups based on a median split, using demographic, psychosocial, and sharenting-related predictors. EWB was positively associated with QoL (r = 0.68, p < .01), social support (r = 0.51, p < .01), and sharenting behavior (r = 0.41, p < .01), representing large, large, and moderate associations, respectively. The ANN was used as a flexible predictive approach capable of accommodating potentially nonlinear and non-additive relationships among predictors. The model achieved classification accuracies of 84.4% in the training set, 82.9% in the testing set, and 75.9% in the holdout set. QoL, sharenting behavior, and social support showed the highest relative importance for EWB classification, with normalized importance values of 100%, 60.8%, and 58.8%, respectively. In contrast, all demographic variables had normalized importance values below 50%. Classification accuracy was higher for mothers in the above-median EWB group than for those in the below-median EWB group. Receiver operating characteristic analysis indicated discrimination above chance. These findings may inform psychoeducational and parent-education programs designed to strengthen mothers' social support, enhance adaptive coping and QoL, and promote reflective, developmentally appropriate, and privacy-conscious digital parenting practices.