Forecasting: principles and practice

By Rob J Hyndman, George Athanasopoulos. 2013

4,399 citations recorded by the source corpus.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Time-Series & Forecasting

Citations in the corpus

Aequitas: A Bias and Fairness Audit Toolkit (2018, methodology): “ing from demographic to behavioral to spatiotemporal aggregations over dierent time and space resolutions) and building several hundred models. We validate our models using temporal cross-validation Hyndman and Athanasopoulos (2014). We split the data into several train-test splits by time, tting mo”
Forecasting Across Time Series Databases using Long Short-Term Memory Networks on Groups of Similar Series (2017, background): “It consists of a sequence of applications of a loess smoother, making the decomposition computationally efficient, even for longer time series (Cleveland et al., 1990; Hyndman and Athanasopoulos, 2014).”
Decision support from financial disclosures with deep neural networks and transfer learning (2017, neutral): “Algorithm 1 describes the tuning in order to ﬁnd the best-performing parameters based on time-series cross-validation that employs a rolling forecasting origin [43].”
Forecast combinations: An over 50-year review (2022, methodology): “Time series cross-validation, also known as “evaluation on a rolling forecasting origin” (Hyndman and Athanasopoulos, 2021), is also recommended in the training procedures for both individual models and meta-models to help with the parameter estimation.”
Workload Prediction Using ARIMA Model and Its Impact on Cloud Applications’ QoS (2015, methodology): “This method is broadly composed of three sequential stages [21]: (i) determination of the number of differencing steps of the model (parameter d); (ii) actual differentiation of the time series d times; and (iii) selection of the best fit model.”
Spatiotemporal modeling and prediction in cellular networks: A big data enabled deep learning approach (2017, background): “In [38], Zhou et al. proposed a network traffic prediction model, which is a combination of linear time series ARIMA model and non-linear GARCH model.”
Evaluating time series forecasting models: an empirical study on performance estimation methods (2019, background): “Essentially, a time series is said to be stationary if its characteristics do not depend on the time that the data is observed [20].”
One Fits All: Power General Time Series Analysis by Pretrained LM (2023, background): “Time series analysis is a fundamental problem Hyndman & Athanasopoulos (2021) that has played an important role in many real-world applications Wen et al. (2022), such as retail sales forecasting Böse et al. (2017); Courty & Li (1999) , imputation of missing data for economic time series Friedman…”
Darts: User-Friendly Modern Machine Learning for Time Series (2021, methodology): “Smoothing often give good results (Hyndman and Athanasopoulos, 2018), machine learning is becoming a more attractive option to improve the models’ representation power and scale to larger datasets and higher dimensionalities.”
Short term electricity demand forecasting using partially linear additive quantile regression with an application to the unit commitment problem (2018, methodology): “For data which exhibits seasonality, [25] propose that the SARIMA(2,0,0)(1,0,0)[s] model be used as an initial model for modeling the residuals in step (1) above.”
On the importance of the long-term seasonal component in day-ahead electricity price forecasting (2016, background): “the stochastic component; Xt), either in an additive or a multiplicative fashion (Hyndman and Athanasopoulos, 2013; Janczura et al., 2013).”
Exploring Interpretable LSTM Neural Networks over Multi-Variable Data (2019, methodology): “ARIMAX is the auto-regressive integrated moving average with regression terms on exogenous variables (Hyndman & Athanasopoulos, 2014).”