Monday, Aug 5: 8:30 AM - 10:20 AM
1091
Invited Paper Session
Oregon Convention Center
Room: CC-B119
Applied
Yes
Main Sponsor
Business and Economic Statistics Section
Co Sponsors
Korean International Statistical Society
Presentations
Managing high-frequency market data has been a challenging task in finance. A limit order book is a collection of orders that a trader intends to place, either to buy or sell at a certain price. Traditional approaches often fall short in forecasting future limit orders because of their high frequency and volume. In this study, we propose a modified attention algorithm to analyze the movement patterns in a limit order book. The enormous amount of data with millisecond time stamps are efficiently examined and processed using an attention module, which highlights important aspects of limit orders. We demonstrate that our modified attention algorithm improves the forecasting accuracy of limit orders.
VAE (Variational Auto Encoder) is a nonlinear factor model which is popularly used in AI tasks
such as generation of synthetic data. VAE models the distribution of given data as the sum of
a nonlinear transformation of latent factors and noise. In most cases, the noise term is assumed to
be Gaussian with the constant variance. In this research, we propose a VAE model where the noise
term also depends on the latent factor, which we call the heterogeneous VAE (H-VAE). We provide motivational
examples for H-VAE, theoretical justification and empirical comparisons with other existing VAE algorithms.
Bayesian neural networks are a powerful tool for characterizing predictive uncertainty, but they face two challenges in practice. First, it is difficult
to define meaningful priors for the weights of the network. Second, conventional computational strategy becomes impractical for large and complex applications. In this paper, we adopt a class of implicit generative priors and propose a novel neural adaptive empirical Bayes framework for Bayesian modeling and inference. These priors are derived through a nonlinear transformation of a known low-dimensional distribution, allowing us to handle complex data distributions and capture the underlying manifold structure effectively. Our framework combines variational inference with a gradient ascent algorithm, which serves to select the hyperparameter and approximate the posterior distribution. Theoretical justification is established through both the posterior and classification consistency. We demonstrate the practical applications of our framework through extensive examples, including two-spiral problem, regression, and 10 UCI datasets, as well as MNIST image classification. The results of our experiments highlight the superiority of o
Recently, a James-Stein shrinkage (JS) estimator has gained attention as a powerful tool for estimating the leading eigenvector of covariance matrices. The efficacy of the JS estimator has been demonstrated under a strongly-spiked leading eigenvalue model, using the high-dimensional, low-sample-size (HDLSS) asymptotic regime, where the number of variables increases while the sample size remains fixed. In this work, we extend the application of the JS shrinkage to the regime of moderately-spiked leading eigenvalues, and reveal a key condition, involving a signal-to-noise ratio, for the JS estimator to be useful. Furthermore, we develop shrinkage estimators for principal component variance and scores, enabling their application in high-dimensional principal component analysis. The implication of the work in Markowitz's mean-variance optimal portfolio will be discussed as well.
We discuss a multiscale volatility analysis of high-frequency intraday prices. Our multiscale framework includes a fractional Brownian motion and microstructure noise as the building blocks. The proposed noisy fractional Brownian motion model is shown to possess a variety of volatility behaviors suitable for intraday price processes. Algorithms for numerical estimation from time series observations are presented with a new Hurst exponent estimator proposed for the noisy fractional Brownian motion model. Using real-world high-frequency price data for a collection of US stocks and ETFs, we estimate the parameters in the noisy fractional Brownian motion and illustrate how the volatility varies over different timescales. The Hurst exponent and noise level exhibit an intraday pattern whereby the noise ratio tends to be higher near market close.
In this work, we model the dynamics of the Best Bid and Ask price of a financial asset using the Hawkes-flocking model and study the optimal placement problem under the model. The Hawkes flocking model is modified version of multivariate Hawkes process with flocking mechanism, derived from a stochastic Cucker-Smale flocking system. Under this price process model, we study the optimal order placement problem: the trader makes its decision at discrete time points until time T to maximize its cash flow under restrictions. We derive the solution under different market regimes.
Speaker
Hyoeun Lee, University of Illinois at Urbana Champaign