Abstract:We discuss Bayesian forecasting of increasingly high-dimensional time series, a key area of application of stochastic dynamic models in the financial industry and allied areas of business. Novel state-space models characterizing sparse patterns of dependence among multiple time series extend existing multivariate volatility models to enable scaling to higher numbers of individual time series. The theory of these "dynamic dependence network" models shows how the individual series can be "decoupled" for sequential analysis, and then "recoupled" for applied forecasting and decision analysis. Decoupling allows fast, efficient analysis of each of the series in individual univariate models that are linked-- for later recoupling-- through a theoretical multivariate volatility structure defined by a sparse underlying graphical model. Computational advances are especially significant in connection with model uncertainty about the sparsity patterns among series that define this graphical model; Bayesian model averaging using discounting of historical information builds substantially on this computational advance. An extensive, detailed case study showcases the use of these models, and the improvements in forecasting and financial portfolio investment decisions that are achievable. Using a long series of daily international currency, stock indices and commodity prices, the case study includes evaluations of multi-day forecasts and Bayesian portfolio analysis with a variety of practical utility functions, as well as comparisons against commodity trading advisor benchmarks.

Forecasting high-dimensional dynamics exploiting suboptimal embeddings

Predicting multiple observations in complex systems through low-dimensional embeddings

DEFM: Delay E mbedding based Forecast Machine for Time Series Forecasting by Spatiotemporal Information Transformation

Randomly Distributed Embedding Making Short-Term High-Dimensional Data Predictable

Forecasting the dynamics of correlations in complex systems

Enhancing Time Series Forecasting: A Hierarchical Transformer with Probabilistic Decomposition Representation

A hybrid framework for multivariate long-sequence time series forecasting

Dynamic dependence networks: Financial time series forecasting and portfolio decisions (with discussion)

Featurizing Koopman Mode Decomposition For Robust Forecasting

High-dimensional Multivariate Time Series Forecasting in IoT Applications using Embedding Non-stationary Fuzzy Time Series

Forecasting causal dynamics with universal reservoirs

Forecasting Outside the Box: Application-Driven Optimal Pointwise Forecasts for Stochastic Optimization

Multiple-Input Multiple-Output Randomized Fuzzy Cognitive Map Method for High-Dimensional Time Series Forecasting

Online Forecasting Matrix Factorization

High-Dimensional Multivariate Forecasting with Low-Rank Gaussian Copula Processes

FDF: Flexible Decoupled Framework for Time Series Forecasting with Conditional Denoising and Polynomial Modeling

Adaptive Graph Convolutional Network Framework for Multidimensional Time Series Prediction

Towards better time series prediction with model-independent, low-dispersion clusters of contextual subsequence embeddings

Forecasting of Nonlinear Dynamics Based on Symbolic Invariance

Dynamic Hypergraph Structure Learning for Multivariate Time Series Forecasting