This is a master thesis project done by Ezgi Günbatar for Applied Data Science program at Utrecht University. The repository contains codes and the codebook which are used in "The Role of Social Networks and Personal Characteristics in Shaping Fertility Intentions: A Multi-Method Machine Learning Perspective" study.
The study aims to find an answer to the “Do the individual characteristics and the social network attributes affect women’s desire to have children?” question. For this purpose, individual and network characteristics are used for predicting the having children intentions.
In this context, methods such as Graph Neural Networks, Gradient Boosting Decision Tree, Random Forest, and Support Vector Machine were used for predicting fertility intentions. Go to the code file to see each method that was used and their results.
The data was collected through the LISS (Longitudinal Internet studies for the Social Sciences) panel, and can be downloaded from https://www.dataarchive.lissdata.nl/study_units/view/1377. The data is freely available, but requires registration. Preprocessed data can't be uploaded here because use of the data requires permission.
For running the R code (preprocessing part) it is important that you download the files "wj18a_EN_1.0p.sav" and "avars_201802_EN_1.0p.sav" to be stored in working directory,downloaded from https://www.lissdata.nl/