Diffusion index forecasting with tensor data
Researchers have introduced a new economic forecasting method that merges tensor data (multi-dimensional arrays) with traditional data, using a Canonical Polyadic tensor factor model. Their framework delivers accurate forecasts even with few non-tensor predictors and includes an analytical formula for prediction intervals that reflects uncertainty in hidden factors. A novel thresholding estimator improves handling of high-dimensional data. Tests on U.S. trade flows show this approach outperforms leading models, though computational challenges remain to be addressed.
What it examines
This paper studies how to improve forecasting using both tensor (multi-dimensional) and regular data. The authors use a special tensor factor model, develop new estimation methods, and analyze prediction accuracy. They also test their approach with simulations and real U.S. trade flow data.
What it concludes
The results show that the proposed methods outperform existing techniques, especially for complex, high-dimensional data. These methods can be used in economics and finance for better forecasting. The study suggests further research on more flexible models and applications to other types of large-scale data.
Evidence objects
Researchers unveil a groundbreaking economic forecasting method that fuses tensor and traditional data, preserving complex structures via a Canonical Polyadic (CP) tensor factor modelan innovation for high-dimensional, multi-source datasets.
key_findings bullet 1 · key_findings · validation V0
Key advances include an analytical formula for prediction intervals that accounts for hidden factor uncertainty, and a novel thresholding estimator for high-dimensional covariance matrices, boosting resilience to cross-sectional dependence in big data.
key_findings bullet 2 · key_findings · validation V0
Simulation and real-world tests on U.S. trade flows show the new approach consistently outperforms popular alternatives, though the authors note future work should address computational complexity and broader application challenges.
key_findings bullet 3 · key_findings · validation V0
This paper pioneers diffusion index forecasting with tensor data, preserving structure via a novel CP tensor factor model. It innovates with robust thresholding for high-dimensional covariance and a multi-source factor-augmented sparse regression. Empirical validation and theoretical advances make it compelling for Economics, Macroeconomics, and Quantitative Finance, despite building on established models.
key_findings bullet 4 · key_findings · validation V0
Raw abstract and provenance
In this paper, we consider diffusion index forecasting with both tensor and non-tensor predictors, where the tensor structure is preserved with a Canonical Polyadic (CP) tensor factor model. When the number of non-tensor predictors is small, we study the asymptotic properties of the least squares estimator in this tensor factor-augmented regression, allowing for factors with different strengths. We derive an analytical formula for prediction intervals that accounts for the estimation uncertainty of the latent factors. In addition, we propose a novel thresholding estimator for the high-dimensional covariance matrix that is robust to cross-sectional dependence. When the number of non-tensor predictors exceeds or diverges with the sample size, we introduce a multi-source factor-augmented sparse regression model and establish the consistency of the corresponding penalized estimator. Simulation studies validate our theoretical results and an empirical application to U.S. trade flows demonstrates the advantages of our approach over other popular methods in the literature.
Source row: 607 · abstract type: unknown