Determinants of Fertility - An Application of Machine Learning Techniques

The paper at hand applies machine learning techniques to investigate first birth transitions. The methods do not rely on distribution assumptions and require only few preconditions for application. The results are compatible with contemporary demographic research, highlighting - among other factors - the status of relationship, income and the distribution of labour in the family. Machine learning techniques may thus be used as explorative method in the social sciences as well as tool for an in-depth analysis in future research as they are especially suited to process large data sets.