Forecasting Big Time Series: Old and New

@article{Faloutsos2018ForecastingBT,
  title={Forecasting Big Time Series: Old and New},
  author={Christos Faloutsos and Jan Gasthaus and Tim Januschowski and Bernie Wang},
  journal={Proc. VLDB Endow.},
  year={2018},
  volume={11},
  pages={2102-2105}
}
Time series forecasting is a key ingredient in the automation and optimization of business processes: in retail, deciding which products to order and where to store them depends on the forecasts of future demand in different regions; in cloud computing, the estimated future usage of services and infrastructure components guides capacity planning; and workforce scheduling in warehouses, call centers, factories requires forecasts of the future workload. Recent years have witnessed a paradigm… 

Forecasting Big Time Series: Theory and Practice

This tutorial provides a concise and intuitive overview of the most important methods and tools available for solving large-scale forecasting problems, and reviews the state of the art in both classical modeling of time series, and deep learning for forecasting.

Classical and Contemporary Approaches to Big Time Series Forecasting

The objective of this tutorial is to provide a concise and intuitive overview of the most important methods and tools available for solving large-scale forecasting problems in three related fields: classical modeling of time series, scalable tensor methods, and deep learning for forecasting.

Deep Learning based Time Series Forecasting

This work has categorized various deep learning-based time series forecasting techniques based on the underlying structures of the technique, such as RNN, CNN, and Transformer, and provided a consolidated report.

Learning Periods from Incomplete Multivariate Time Series

A robust framework for multivariate period detection, called LAPIS, which encodes incomplete and noisy data as a sparse summary via a Ramanujan periodic dictionary and enables improvements in downstream applications such as forecasting, missing value imputation and clustering.

Koopman Neural Forecaster for Time Series with Temporal Distribution Shifts

A novel deep sequence model based on the Koopman theory for time series forecasting: KNF that leverages DNNs to learn the linear Koop man space and the coefficients of chosen measurement functions, achieves SOTA prediction performance on M4, Cryptos and NBA player trajectory datasets and provides interpretable results.

Deep Factors for Forecasting

A hybrid model that incorporates the benefits of both classical and deep neural networks is proposed, which is data-driven and scalable via a latent, global, deep component, and handles uncertainty through a local classical model.

Intermittent Demand Forecasting with Deep Renewal Processes

A new, unified framework for building intermittent demand forecasting models is introduced, which incorporates and allows to generalize existing methods in several directions, based on extensions of well-established model-based methods to discrete-time renewal processes, which can parsimoniously account for patterns such as aging, clustering and quasi-periodicity in demand arrivals.

Transforming Autoregression: Interpretable and Expressive Time Series Forecast

Autoregressive Transformation Models (ATMs), a model class inspired from various research directions such as normalizing flows and autoregressive models that unite expressive distributional forecasts using a semi-parametric distribution assumption with an interpretable model specification and allow for uncertainty quantification based on (asymptotic) Maximum Likelihood theory are proposed.

Deep Learning for Time Series Forecasting: Tutorial and Literature Survey

An introduction and overview of the field is provided and important building blocks for deep forecasting in some depth are presented; using these building blocks, the breadth of the recent deep forecasting literature is surveyed.

References

SHOWING 1-10 OF 47 REFERENCES

Probabilistic Demand Forecasting at Scale

A platform built on large-scale, data-centric machine learning approaches, whose particular focus is demand forecasting in retail, that enables the training and application of probabilistic demand forecasting models, and provides convenient abstractions and support functionality for forecasting problems.

DeepAR: Probabilistic Forecasting with Autoregressive Recurrent Networks

Fast mining and forecasting of complex time-stamped events

TriMine is introduced, which performs three-way mining for all three attributes, namely, URLs, users, and time, and consistently outperforms the best state-of-the-art existing methods in terms of accuracy and execution speed.

Query-based Workload Forecasting for Self-Driving Database Management Systems

This work presents a robust forecasting framework called QueryBot 5000 that allows a DBMS to predict the expected arrival rate of queries in the future based on historical data and presents a clustering-based technique for reducing the total number of forecasting models to maintain.

Online data mining for co-evolving time sequences

This work develops a fast method to analyze co-evolving time sequences jointly to allow estimation/forecasting of missing/delayed/future values, quantitative data mining, and outlier detection, and adapts to changing correlations among time sequences.

A Multi-Horizon Quantile Recurrent Forecaster

We propose a framework for general probabilistic multi-step time series regression. Specifically, we exploit the expressiveness and temporal nature of Recurrent Neural Networks, the nonparametric

ARMDN: Associative and Recurrent Mixture Density Networks for eRetail Demand Forecasting

A Neural Network architecture called AR-MDN, that simultaneously models associative factors, time-series trends and the variance in the demand, is proposed, that yields a significant improvement in forecasting accuracy when compared with existing alternatives.

Temporal Regularized Matrix Factorization for High-dimensional Time Series Prediction

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.

Predicting the Present with Bayesian Structural Time Series

The system combines a structural time series model for the target series with regression component capturing the contributions of contemporaneous search query data, and averages over potential contributions from a very large set of models.

Time series analysis, forecasting and control

time series analysis san francisco state university, 6 4 introduction to time series analysis, box and jenkins time series analysis forecasting and, th15 weeks citation classic eugene garfield, proc