Identifying the Stability of Couple Relationship Applying Different Machine Learning Techniques

A couple relationship is an important matter where two people engaged each other in a romantic association. This bonding happens almost all human beings throughout the world. The failure of a relationship is a serious issue that affects not only two directly involved people but also hampered the kinship of indirectly associated people. This work aims to investigate and estimate the persistence of the couple relationship by using machine learning models. Therefore, we gathered a couple relationship data from the Standford Social Science Data Collection repository and reprocessed them for further analysis. Then, we generated several feature subsets from the primary dataset and employed different machine learning classifiers to them. In this case, the Bagging classifier with a decision tree classifier shows the best accuracy of 87% for the recursive feature elimination feature subset. We also explored significant features from these subsets which can be used to estimate the stability of a couple relationship as well. Thus, our model can be useful to predict in advance about the future relationship and could be possible to recover them before any occurrence.