Corpus ID: 231933997

Tight Risk Bound for High Dimensional Time Series Completion

@inproceedings{Alquier2021TightRB,
  title={Tight Risk Bound for High Dimensional Time Series Completion},
  author={Pierre Alquier and Nicolas Marie and Am'elie Rosier},
  year={2021}
}
Initially designed for independent datas, low-rank matrix completion was successfully applied in many domains to the reconstruction of partially observed high-dimensional time series. However, there is a lack of theory to support the application of these methods to dependent datas. In this paper, we propose a general model for multivariate, partially observed time series. We show that the least-square method with a rank penalty leads to reconstruction error of the same order as for independent… 

Figures and Tables from this paper

References

SHOWING 1-10 OF 47 REFERENCES
Structured low-rank matrix completion for forecasting in time series analysis
TLDR
This paper considers a matrix completion problem for Hankel matrices and a convex relaxation based on the nuclear norm and investigates the cases when the proposed approach can work.
Temporal Regularized Matrix Factorization for High-dimensional Time Series Prediction
TLDR
This paper develops novel regularization schemes and uses scalable matrix factorization methods that are eminently suited for high-dimensional time series data that has many missing values, and makes interesting connections to graph regularization methods in the context of learning the dependencies in an autoregressive framework.
Low Rank and Structured Modeling of High-Dimensional Vector Autoregressions
TLDR
This work introduces a novel approach for estimating low-rank and structured sparse high-dimensional VAR models using a regularized framework involving a combination of nuclear norm and lasso penalties, and establishes nonasymptotic probabilistic upper bounds on the estimation error rates of the low- rank and the structured sparse components.
Estimation of latent factors for high-dimensional time series
This paper deals with the dimension reduction of high-dimensional time series based on a lower-dimensional factor process. In particular, we allow the dimension of time series N to be as large as, or
Temporal Dynamic Matrix Factorization for Missing Data Prediction in Large Scale Coevolving Time Series
TLDR
The approaches are optimally designed to largely utilize both the interior patterns of each time series and the information of time series across multiple sources to build an initial model, and imposed hybrid regularization terms to constrain the objective functions of matrix factorization.
High-dimensional VAR with low-rank transition
We propose a vector auto-regressive model with a low-rank constraint on the transition matrix. This model is well suited to predict high-dimensional series that are highly correlated, or that are
Probabilistic low-rank matrix completion on finite alphabets
TLDR
This work investigates the case where the observations take a finite number of values, corresponding for examples to ratings in recommender systems or labels in multi-class classification, and derives bounds for the Kullback-Leibler divergence between the true and estimated distributions.
Singular spectrum-based matrix completion for time series recovery and prediction
TLDR
This work proposes Singular Spectrum Matrix Completion (SS-MC), a novel approach for the simultaneous recovery of missing data and the prediction of future behavior in the absence of complete measurement sets, achieving lower estimation error compared to state-of-the-art schemes.
Nonnegative Matrix Factorization with Side Information for Time Series Recovery and Prediction
TLDR
A new algorithm (HALSX, or Hierarchical Alternating Least Squares with eXogeneous variables) is proposed which estimates NMF in this setting and is validated on both simulated and real electricity consumption datasets as well as a recommendation system dataset.
Factor modeling for high-dimensional time series: inference for the number of factors
This paper deals with the factor modeling for high-dimensional time series based on a dimension-reduction viewpoint. Under stationary settings, the inference is simple in the sense that both the
...
1
2
3
4
5
...