使用具有非常大的数据集的广义加法模型编码预测区间

I have a small dataset of people, thier locations, and whether or not they know each other. It is a subset of a dataset with 1000 people. Given that each person could know any other person, the number of potential links grows just under n^2. I want to fit a model with the small subset, to get probability of linkage as a function of distance, so that I can perform simulations with the wider dataset. I have a small dataset of people, thier locatio