The empirical Christoffel function with applications in data analysis (Q2000530)

After the definition and discussion of the properties of the Christoffel function, the empirical Christoffel function is introduced, which has an input measure \(\mu\) being a scaled counting measure uniformly supported on a set of data points. The aim of the article is then to present application of the empirical Christoffel function to machine learning procedures. The function can be used also for density estimation, support inference (recovering the unknown support of a population measure \(\mu\) from \(n\) independent samples) and outlier detection. Several numerical applications in statistics and machine learning are treated.