Interpreting Machine Learning Model for Generic Needs on Marital Satisfaction

Marital relationship quality is vital to happiness and mental or physical health of the couple and family. A highly satisfying marriage is highly resistant to union dissolution and family fragmentation. Marital satisfaction, a measurement of the couple relationship, is not only important for self-evaluation for the healthy relationship of the couple, but also as aggregate measurement for the health and sustainability of society as well. Healthy family enables Malaysia to attain Sustainable Development Goals of SDG 1, 2, 3, 4, 5, 8, 10, 11, 12, 16 & 17. The negative economic and social impacts of divorce on women and especially children have been well-documented. Many of the social problems of the youth are a result of the youth grow up without the benefits of two caring parents. Collectively, high divorce rate poses a serious risk for the sustainable development of society. As society progresses, unfortunately, Malaysia also appears futile to resist the high level in divorce rates. Almost 80000 of divorce cases were filed during the first and half year of the pandemic period nationally. This poses a serious and urgent threat to Malaysia ' s National Family Policy and highlights inadequacies of previous studies that warrant additional research. Many proposed solutions are trying to solve the phenomenal symptoms or pertaining to a certain culture that are not applicable across diverse ethnics in Malaysia. Understanding cross-cultural predictors for marital satisfaction is difficult. Various researchers show inconsistent predictors and varying results across machine learning works on marital satisfaction. Using interpreting machine learning on cross-cultural data from 33 Countries (N= 7,767), the objective is to identify a model that able to interpret the nature, dynamics and intricacies of combined generic needs on marital satisfaction. A variety of explanatory theories in the domains of happiness and marital satisfaction will be discussed. Finally, a new machine learning marital satisfaction prediction algorithm on generic needs will be formulated, and a new framework that offers a better interpretability of generic needs input and satisfaction output will be developed. Understanding the impacts of generic needs on satisfying relationship is expected to provide better insights for couple's decision-making, government's policy recommendation, future research in cross-cultural universal values and educational program to enhance marital satisfaction and reduce divorces due to incompatibility and infidelity