Pietro Coretto

In cluster analysis one often finds several partitions of a data set using different clustering methods and algorithms set with a variety of hyperparameters and tunings. The number of clusters K is one of the most relevant of such hyperparameters. Cluster selection is the task of choosing the desired partitions. The Bootstrap Quadratic Scoring is a recently introduced method where the cluster selection is performed by optimizing a score attached to a partition that is based on the quadratic discriminant function. Previously, we proposed the estimation of this cluster score via bootstrap resampling and investigated the proposed estimator based on numerical experiments and real data applications. However, that earlier work did not provide theoretical guarantees. In this paper, we fill that gap. We study the asymptotic behavior of the scoring method and show that the proposed estimator converges to well-defined population counterparts.

Summary: We introduce a new approach to deciding the number of clusters. The approach is applied to Optimally Tuned Robust Improper Maximum Likelihood Estimation (OTRIMLE; Coretto & Hennig, Journal of the American Statistical Association 111, 1648–1659) of a Gaussian mixture model allowing for observations to be classified as ‘noise’, but it can be applied to other clustering methods as well. The quality of a clustering is assessed by a statistic Q that measures how close the within‐cluster distributions are to elliptical unimodal distributions that have the only mode in the mean. This non‐parametric measure allows for non‐Gaussian clusters as long as they have a good quality according to Q. The simplicity of a model is assessed by a measure S that prefers a smaller number of clusters unless additional clusters can reduce the estimated noise proportion substantially. The simplest model is then chosen that is adequate for the data in the sense that its observed value of Q is not significantly larger than what is expected for data truly generated from the fitted model, as can be assessed by parametric bootstrap. The approach is compared with model‐based clustering using the Bayesian information criterion (BIC) and the integrated complete likelihood (ICL) in a simulation study and on two real data sets.

In this paper we study a finite Gaussian mixture model with an additional uniform component that has the role to catch points in the tails of the data distribution. An adaptive constraint enforces a certain level of separation between the Gaussian mixture components and the uniform component representing noise and outliers in the tail of the distribution. The latter makes the proposed tool particularly useful for robust estimation and outlier identification. A constrained ML estimator is introduced for which existence and consistency is shown. One of the attractive features of the methodology is that the noise level is estimated from data. We also develop an EM-type algorithm with proven convergence. Based on numerical evidence we show how the methods developed in this paper are useful for several fundamental data analysis tasks: outlier identification, robust location-scale estimation, clustering, and density estimation.

The inhomogeneity of the cross-sectional distribution of realized assets’ volatility is explored and used to build a novel class of GARCH (Generalized Autoregressive Conditional Heteroskedasticity) models. The inhomogeneity of the cross-sectional distribution of realized volatility is captured by a finite Gaussian mixture model plus a uniform component that represents abnormal variations in volatility. Based on the cross-sectional mixture model, at each time point, memberships of assets to risk groups are retrieved via maximum likelihood estimation, as well as the probability that an asset belongs to a specific risk group. The latter is profitably used for specifying a state-dependent model for volatility forecasting. We propose novel GARCH-type specifications the parameters of which act “clusterwise” conditional on past information on the volatility clusters. The empirical performance of the proposed models is assessed by means of an application to a panel of U.S. stocks traded on the NYSE. An extensive forecasting experiment shows that, when the main goal is to improve overall many univariate volatility forecasts, the method proposed in this paper has some advantages over the state-of-the-arts methods.