Heat Kernel Analysis of Syntactic Structures

We consider two different data sets of syntactic parameters and we discuss how to detect relations between through a heat kernel method developed by Belkin–Niyogi, which produces low dimensional representations the data, based on Laplace eigenfunctions, that preserve neighborhood information. analyze connectivity clustering structures arise in datasets, regions maximal variance two-parameter space Belkin–Niyogi construction, identify preferable choices independent variables. compute coefficients their variance.