Claudia Tarantola

Monitoring and improving Greek banking services using Bayesian Networks: An analysis of mystery shopping data

Mystery shopping is a well known marketing technique used by companies and marketing analysts to measure quality of service, and gather information about products and services. In this article, we analyse data from mystery shopping surveys via Bayesian Networks in order to examine and evaluate the quality of service offered by the loan departments of Greek Banks. We use mystery shopping visits to collect information about loan products and services and, by this way, evaluate the customer satisfaction and plan improvement strategies that will assist banks to reach their internal standards. Bayesian Networks not only provide a pictorial representation of the dependence structure between the characteristics of interest but also allow to evaluate, interpret and understand the effects of possible improvement strategies.

Default Probability Estimation via Pair Copula Constructions

In this paper we present a novel approach for firm default probability estimation. The methodology is based on multivariate contingent claim analysis and pair copula constructions. For each considered firm, balance sheet data are used to assess the asset value, and to compute its default probability. The asset pricing function is expressed via a pair copula construction, and it is approximated via Monte Carlo simulations. The methodology is illustrated through an application to the analysis of both operative and defaulted firms.

MCMC model determination for discrete graphical models

In this paper we compare two alternative MCMC samplers for the Bayesian analysis of discrete graphical models; we present both a hierarchical and a nonhierarchical version of them. We first consider the MC 3 algorithm by Madigan and York (1995) for which we propose an extension that allows for a hierarchical prior on the cell counts. We then describe a novel methodology based on a reversible jump sampler. As a prior distribution we assign, for each given graph, a hyper-Dirichlet distribution on the matrix of cell probabilities. Two applications to real data are presented.

Body plethysmographic study of specific airway resistance in a sample of healthy adults.

Background and objective:  sRaw (specific airway resistance) is a corrected index (Raw multiplied by thoracic gas volume) that describes airway behaviour regardless of lung volume. Normal values of sRaw in adult subjects have never been formally defined. To establish sRaw interpretation criteria and to define a range of reference values, we evaluated variability, reproducibility and reliability of sRaw measurements in a group of healthy adults.

Bayesian Value-at-Risk with product partition models

In this paper we propose a novel Bayesian methodology for Value-at-Risk computation based on parametric Product Partition Models. Value-at-Risk is a standard tool for measuring and controlling the market risk of an asset or portfolio, and is also required for regulatory purposes. Its popularity is partly due to the fact that it is an easily understood measure of risk. The use of Product Partition Models allows us to remain in a Normal setting even in the presence of outlying points, and to obtain a closed-form expression for Value-at-Risk computation. We present and compare two different scenarios: a product partition structure on the vector of means and a product partition structure on the vector of variances. We apply our methodology to an Italian stock market data set from Mib30. The numerical results clearly show that Product Partition Models can be successfully exploited in order to quantify market risk exposure. The obtained Value-at-Risk estimates are in full agreement with Maximum Likelihood approaches, but our methodology provides richer information about the clustering structure of the data and the presence of outlying points.

Bayesian outlier detection in Capital Asset Pricing Model

We propose a novel Bayesian optimization procedure for outlier detection in the Capital Asset Pricing Model. We use a parametric product partition model to robustly estimate the systematic risk of an asset. We assume that the returns follow independent normal distributions and we impose a partition structure on the parameters of interest. The partition structure imposed on the parameters induces a corresponding clustering of the returns. We identify via an optimization procedure the partition that best separates standard observations from the atypical ones. The methodology is illustrated with reference to a real dataset, for which we also provide a microeconomic interpretation of the detected outliers.

GLOBAL PRIOR DISTRIBUTIONS FOR THE ANALYSIS OF DISCRETE GRAPHICAL MODELS

We propose a new class of prior distributions for the analysis of discrete graphical models. Such a class, obtained following a conditional approach, generalizes the hyper Dirichlet distributions of Dawid and Lauritzen (1993), since it can be extended to non decomposable graphical models. The two classes are compared in terms of model selection, with an application to a medical data-set illustrating the performance of the two resulting procedures. The proposed class turns out to select simpler, more par-simonious structures.

Multivariate dependence analysis via tree copula models: An application to one-year forward energy contracts

Abstract We propose a novel multivariate approach for dependence analysis in the energy market. The methodology is based on tree copulas and GARCH type processes. We use it to study the dependence structure among the main factors affecting energy price, and to perform portfolio risk evaluation. The temporal dynamic of the examined variables is described via a set of GARCH type models where the joint distribution of the standardised residuals is represented via suitable tree copula structures. Working in a Bayesian framework, we perform both qualitative and quantitative learning. Posterior summaries of the quantities of interest are obtained via MCMC methods.

Spanning trees and identifiability of a single-factor model

The aim of this paper is to propose conditions for exploring the class of identifiable Gaussian models with one latent variable. In particular, we focus attention on the topological structure of the complementary graph of the residuals. These conditions are mainly based on the presence of odd cycles and bridge edges in the complementary graph. We propose to use the spanning tree representation of the graph and the associated matrix of fundamental cycles. In this way it is possible to obtain an algorithm able to establish in advance whether modifying the graph corresponding to an identifiable model, the resulting graph still denotes identifiability.

Bayesian Networks for Financial Market Signals Detection

In order to model and explain the dynamics of the market, different types and sources of information should be taken into account. We propose to use a Bayesian network as a quantitative financial tool for market signals detection. We combine and incorporate in the model, accounting, market, and sentiment data. The network is used to describe the relationships among the examined variables in an immediate way. Furthermore, it permits to identify in a mouse-click time scenario that could lead to operative signals. An application to the analysis of S&P 500 index is presented.