Penalized cluster analysis with applications to family data

Cluster analysis is the assignment of observations into clusters so that observations in the same cluster are similar in some sense, and many clustering methods have been developed. However, these methods cannot be applied to family data, which possess intrinsic familial structure. To take the familial structure into account, we propose a form of penalized cluster analysis with a tuning parameter controlling its influence. The tuning parameter can be selected based on the concept of clustering stability. The method can also be applied to other cluster data such as panel data. The method is illustrated via simulations and an application to a family study of asthma.