The volatility forecasting task refers to predicting the amount of variability in price a financial asset over certain period. It is an important mechanism for evaluating risk associated with and, as such, significant theoretical and practical importance analysis. While classical approaches have framed this time-series prediction one – using historical pricing guide future recent advances natural language processing seen researchers turn complementary sources data, such analyst reports,...

Corporate mergers and acquisitions (M&A) account for billions of dollars investment globally every year offer an interesting challenging domain artificial intelligence. However, in these highly sensitive domains, it is crucial to not only have a robust/accurate model, but be able generate useful explanations garner user's trust the automated system. Regrettably, recent research regarding eXplainable AI (XAI) financial text classification has received little no attention, many current methods...

Spherical regression, where both covariate and response variables are defined on the sphere, is a required form of data analysis in several scientific disciplines, has been subject substantial methodological development recent years. Yet, it remains challenging problem due to complexities involved constructing valid expressive regression models between spherical domains, difficulties quantifying uncertainty estimated maps. To address these challenges, we propose casting as optimal transport...

Spatial processes with nonstationary and anisotropic covariance structure are often used when modeling, analyzing, predicting complex environmental phenomena. Such may be expressed as ones that have stationary isotropic on a warped spatial domain. However, the warping function is generally difficult to fit not constrained injective, resulting in "space-folding." Here, we propose modeling an injective through composition of multiple elemental functions deep-learning framework. We consider two...

Many existing statistical and machine learning tools for social network analysis focus on a single level of analysis. Methods designed clustering optimize global partition the graph, whereas projection-based approaches (e.g., latent space model in statistics literature) represent rich detail roles individuals. pertinent questions sociology economics, however, span multiple scales Further, many involve comparisons across disconnected graphs that will, inevitably be different sizes, either due...

Abstract We propose a weighted stochastic block model (WSBM) which extends the to important case in edges are weighted. address parameter estimation of WSBM by use maximum likelihood and variational approaches, establish consistency these estimators. The problem choosing number classes is addressed. proposed applied simulated data an illustrative set.

Dáil Éireann is the principal chamber of Irish parliament. The 31st was in session from March 11th, 2011 to February 6th, 2016. Many members were active on social media, and many Twitter users who followed other Dáil. pattern following amongst these politicians provides insights into political alignment within We propose a new model, called generalized latent space stochastic blockmodel, which extends generalizes both model blockmodel study media connections between probability an edge two...

A probabilistic model for random hypergraphs is introduced to represent unary, binary and higher order interactions among objects in real-world problems. This an extension of the latent class analysis that introduces two clustering structures hyperedges captures variation size hyperedges. An expectation maximization algorithm with minorization steps developed perform parameter estimation. Model selection using Bayesian Information Criterion proposed. The applied simulated data sets where...

Non-homogeneous Poisson processes are used in a wide range of scientific disciplines, ranging from the environmental sciences to health sciences. Often, central object interest point process is underlying intensity function. Here, we present general model for function non-homogeneous using measure transport. The built flexible bijective mapping that maps simpler reference We enforce bijectivity by modeling map as composition multiple have increasing triangular structure, and show exhibits an...

Abstract The von Mises–Fisher distribution is one of the most widely used probability distributions to describe directional data. Finite mixtures have found numerous applications. However, likelihood function for finite mixture unbounded and consequently maximum estimation not well defined. To address problem degeneracy, we consider a penalized approach whereby penalty incorporated. We prove strong consistency resulting estimator. An Expectation–Maximization algorithm developed experiments...

A probabilistic model for random hypergraphs is introduced to represent unary, binary and higher order interactions among objects in real-world problems. This an extension of the Latent Class Analysis model, which captures clustering structures objects. An EM (expectation maximization) algorithm with MM (minorization steps developed perform parameter estimation while a cross validated likelihood approach employed selection. The applied three data sets where interesting results are obtained.

Abstract Recent work on point processes includes studying posterior convergence rates of estimating a continuous intensity function. In this article, for the function and change‐point are derived more general case piecewise We study problem an inhomogeneous Poisson process with using non‐parametric Bayesian methods. An Markov Chain Monte Carlo (MCMC) algorithm is proposed to obtain estimates which illustrated simulation studies applications. The Canadian Journal Statistics 47: 604–618; 2019...

Corporate mergers and acquisitions (M&A) account for billions of dollars investment globally every year, offer an interesting challenging domain artificial intelligence. However, in these highly sensitive domains, it is crucial to not only have a robust accurate model, but be able generate useful explanations garner user's trust the automated system. Regrettably, recent research regarding eXplainable AI (XAI) financial text classification has received little no attention, many current...

Non-homogeneous Poisson processes are used in a wide range of scientific disciplines, ranging from the environmental sciences to health sciences. Often, central object interest point process is underlying intensity function. Here, we present general model for function non-homogeneous using measure transport. The built flexible bijective mapping that maps simpler reference We enforce bijectivity by modeling map as composition multiple simple maps, and show exhibits an important approximation...

Many existing statistical and machine learning tools for social network analysis focus on a single level of analysis. Methods designed clustering optimize global partition the graph, whereas projection based approaches (e.g. latent space model in statistics literature) represent rich detail roles individuals. pertinent questions sociology economics, however, span multiple scales Further, many involve comparisons across disconnected graphs that will, inevitably be different sizes, either due...