arXiv research

Characterizes statistical complexity of realizable regression in PAC and online learning.

In this paper we formulate a regression problem to predict realized volatility by using option price data and enhance VIX-styled volatility indices' predictability and liquidity. We test algorithms including regularized regression and machine learning methods such as Feedforward Neural Networks (FNN) on S&P 500 Index a…

New bandit algorithm works without realizability assumption.

A new realized conditional autoregressive Value-at-Risk (VaR) framework is proposed, through incorporating a measurement equation into the original quantile regression model. The framework is further extended by employing various Expected Shortfall (ES) components, to jointly estimate and forecast VaR and ES. The measu…

Faster algorithm reduces contextual bandit regret with fewer offline regression calls.

The joint Value at Risk (VaR) and expected shortfall (ES) quantile regression model of Taylor (2017) is extended via incorporating a realized measure, to drive the tail risk dynamics, as a potentially more efficient driver than daily returns. Both a maximum likelihood and an adaptive Bayesian Markov Chain Monte Carlo m…

Study bounds noise level in linear regression with dependent data.

Optimal algorithm for maximizing rewards in contextual bandits with resource constraints.

We study distributions of realized variance (squared realized volatility) and squared implied volatility, as represented by VIX and VXO indices. We find that Generalized Beta distribution provide the best fits. These fits are much more accurate for realized variance than for squared VIX and VXO -- possibly another indi…

This paper investigates how the conditional quantiles of future returns and volatility of financial assets vary with various measures of ex-post variation in asset prices as well as option-implied volatility. We work in the flexible quantile regression framework and rely on recently developed model-free measures of int…

Bayesian framework forecasts financial tail risks using realized volatility and nonlinear thresholds.

Paper develops neural network for distribution regression.

Paper tackles MLR prediction error without assuming realizable models.

This study improves tail risk forecasting by integrating overnight information into semi-parametric models.

New active learning framework for multiclass classification beyond realizability assumption.

In this work, we highlight a connection between the incremental proximal method and stochastic filters. We begin by showing that the proximal operators coincide, and hence can be realized with, Bayes updates. We give the explicit form of the updates for the linear regression problem and show that there is a one-to-one …

New method tests independence with single nonstationary time series.

A major challenge in contextual bandits is to design general-purpose algorithms that are both practically useful and theoretically well-founded. We present a new technique that has the empirical and computational advantages of realizability-based approaches combined with the flexibility of agnostic methods. Our algorit…

Characterizes the sample complexity of list regression tasks.

New ensemble SVM model reduces prediction error without choosing best kernel.

New bounds for agnostic learning with average smoothness.

CAVI speeds up Bayesian MIDAS regression by 107x-1,772x with similar accuracy.

Unified framework for realizable and agnostic learning.

This paper investigates how realized and option implied volatilities are related to the future quantiles of commodity returns. Whereas realized volatility measures ex-post uncertainty, volatility implied by option prices reveals the market's expectation and is often used as an ex-ante measure of the investor sentiment.…

Enhances Gaussian process regression with multi-fidelity models and active subspaces for high-dimensional problems.

A fundamental challenge in contextual bandits is to develop flexible, general-purpose algorithms with computational requirements no worse than classical supervised learning tasks such as classification and regression. Algorithms based on regression have shown promising empirical success, but theoretical guarantees have…

Adaptive sparseness enhances robust regression using MCC and ARD.

New algorithms achieve near-optimal cumulative loss in nonparametric online learning and games.

Bayesian framework improves robustness in nonlinear regression models.

Study on ReLU regression with Massart noise, achieving exact parameter recovery.

This paper presents the nonparametric inference for nonlinear volatility functionals of general multivariate Itô semimartingales, in high-frequency and noisy setting. Pre-averaging and truncation enable simultaneous handling of noise and jumps. Second-order expansion reveals explicit biases and a pathway to bias correc…

This paper revisits the fractional cointegrating relationship between ex-ante implied volatility and ex-post realized volatility. We argue that the concept of corridor implied volatility (CIV) should be used instead of the popular model-free option-implied volatility (MFIV) when assessing the fractional cointegrating r…

Study examines how imputation accuracy affects prediction accuracy in regression problems with missing covariates.

Algorithm identifies bilinear dynamical systems from noisy data.

The paper analyzes the performance of empirical risk minimization for $p$-norm linear regression.

Study highlights how model choice affects uncertainty estimation in neural network regression.

Personalized medicine seeks to identify the causal effect of treatment for a particular patient as opposed to a clinical population at large. Most investigators estimate such personalized treatment effects by regressing the outcome of a randomized clinical trial (RCT) on patient covariates. The realized value of the ou…

The study uses reproducing kernels to model bond discount curves.

The goal of regression analysis is to predict the value of a numeric outcome variable y given a vector of joint values of other (predictor) variables x. Usually a particular x-vector does not specify a repeatable value for y, but rather a probability distribution of possible y--values, p(y|x). This distribution has a l…

In this work, we propose a new Gaussian process regression (GPR) method: physics information aided Kriging (PhIK). In the standard data-driven Kriging, the unknown function of interest is usually treated as a Gaussian process with assumed stationary covariance with hyperparameters estimated from data. In PhIK, we compu…

New method calibrates probabilistic regression models without restrictive assumptions.

Selecting important features in non-linear or kernel spaces is a difficult challenge in both classification and regression problems. When many of the features are irrelevant, kernel methods such as the support vector machine and kernel ridge regression can sometimes perform poorly. We propose weighting the features wit…

This paper proposed a new regression model called $l_1$-regularized outlier isolation and regression (LOIRE) and a fast algorithm based on block coordinate descent to solve this model. Besides, assuming outliers are gross errors following a Bernoulli process, this paper also presented a Bernoulli estimate model which, …

The paper proposes a mixed-frequency quantile regression model for VaR and ES forecasting.

Paper analyzes agnostic learning of mixed linear regression without generative models.

Realized statistics based on high frequency returns have become very popular in financial economics. In recent years, different non-parametric estimators of the variation of a log-price process have appeared. These were developed by many authors and were motivated by the existence of complete records of price data. Amo…

We introduce a concept of $autoregressive$ (AR)state-space realization that could be applied to all transfer functions $\boldsymbol{T}(L)$ with $\boldsymbol{T}(0)$ invertible. We show that a theorem of Kalman implies each Vector Autoregressive model (with exogenous variables) has a minimal $AR$-state-space realization …

New GLS estimator handles high-dimensional data with autocorrelated errors.