Forecasting: theory and practice

Petropoulos, Fotios; Apiletti, Daniele; Assimakopoulos, Vassilios; Babaï, Mohamed Zied; Barrow, Devon K.; Taieb, Souhaib Ben; Bergmeir, Christoph; Bessa, Ricardo J.; Bijak, Jakub; Boylan, John E.; Browell, Jethro; Carnevale, Claudio; Castle, Jennifer L.; Cirillo, Pasquale; Padgett, Carol; Cordeiro, Clara; Oliveira, Fernando Luiz Cyrino; Baets, Shari De; Dokumentov, Alexander; Ellison, Joanne; Fiszeder, Piotr; Franses, Philip Hans; Frazier, David T.; Gilliland, Michael; Gönül, M. Sinan; Goodwin, Paul; Grossi, Luigi; Grushka‐Cockayne, Yael; Guidolin, Mariangela; Guidolin, Massimo; Gunter, Ulrich; Guo, Xiaojia; Guseo, Renato; Harvey, Nigel; Hendry, David F.; Hollyman, Ross; Januschowski, Tim; Jeon, Junho; Jose, Victor Richmond R.; Kang, Yanfei; Koehler, Anne; Kolassa, Stephan; Kourentzes, Nikolaos; Leva, Sonia; Li, Feng; Litsiou, Konstantia; Makridakis, Spyros; Martin, Gael M.; Martinez, Andrew B.; Meeran, Sheik; Modis, Theodore; Νικολόπουλος, Κωνσταντίνος; Önkal, Dilek; Paccagnini, Alessia; Panagiotelis, Anastasios; Panapakidis, Ioannis P.; Pavía, José M.; Pedio, Manuela; Pedregal, Diego J.; Pinson, Pierre; Ramos, Patrícia; Rapach, David E.; Reade, J. James; Rostami-Tabar, Bahman; Rubaszek, Michał; Sermpinis, Georgios; Shang, Han Lin; Spiliotis, Evangelos; Syntetos, Aris A.; Talagala, Priyanga Dilini; Talagala, Thiyanga S.; Tashman, Len; Thomakos, Dimitrios D.; Thorarinsdottir, Thordis L.; Todini, E.; Trapero, Juan R.; Wang, Xiaoqian; Winkler, Robert L.; Yusupova, Alisa; Ziel, Florian

doi:10.1016/j.ijforecast.2021.11.001

Cited by 397 publications

(169 citation statements)

References 1,547 publications

Supporting

Mentioning

101

Contrasting

Unclassified

Order By: Relevance

“…Over the years a number of authors have criticized the exclusive use of statistical error measures to evaluate and compare forecasts. However, a standardized test ground/procedure for evaluating the economic impact of predictions has not been developed, not only in EPF (Hong et al, 2020), but in forecasting in general (Petropoulos et al, 2022). And this, despite the fact that already three decades ago Murphy (1993) postulated that the "goodness" of a forecast can be assessed in terms of consistency, quality, and value.…”

Section: Trend #3: From Statistical To Economic Evaluationmentioning

confidence: 99%

Forecasting Electricity Prices

Maciejowska¹,

Uniejewski²,

Weron³

2022

Preprint

View full text Add to dashboard Cite

Forecasting electricity prices is a challenging task and an active area of research since the 1990s and the deregulation of the traditionally monopolistic and government-controlled power sectors. It is interdisciplinary by nature. It requires expertise in econometrics, statistics or machine learning for developing well-performing predictive models, in finance for understanding market mechanics, and in electrical engineering for comprehension of the fundamentals driving electricity prices.Although electricity price forecasting aims at predicting both spot and forward prices, the vast majority of research is focused on short-term horizons which exhibit dynamics unlike in any other market. The reason is that power system stability calls for a constant balance between production and consumption, while being weather (both demand and supply) and business activity (demand only) dependent. The recent market innovations do not help in this respect. The rapid expansion of intermittent renewable energy sources is not offset by the costly increase of electricity storage capacities and modernization of the grid infrastructure.On the methodological side, this leads to three visible trends in electricity price forecasting research as of 2022. Firstly, there is a slow, but more noticeable with every year, tendency to consider not only point but also probabilistic (interval, density) or even path (also called ensemble) forecasts. Secondly, there is a clear shift from the relatively parsimonious econometric (or statistical) models towards more complex and harder to comprehend, but more versatile and eventually more accurate statistical/machine learning approaches. Thirdly, statistical error measures are nowadays regarded as only the first evaluation step. Since they may not necessarily reflect the economic value of reducing prediction errors, more and more often, they are complemented by case studies comparing profits from scheduling or trading strategies based on price forecasts obtained from different models.

show abstract

Section: Trend #3: From Statistical To Economic Evaluationmentioning

confidence: 99%

Forecasting Electricity Prices

Maciejowska¹,

Uniejewski²,

Weron³

2022

Preprint

View full text Add to dashboard Cite

show abstract

“…where M is the number of weak learners (soft decision trees), D(m) is the set of descendants of the internal node m, LD(m) is the set of left descendants of the internal node m, i.e., it consists of the nodes in the sub-tree rooted at node m which are reached from m by first selecting the left branch of m. (7) and (8) give the necessary components for the update equations of the learnable parameters of the sGBDT, namely, w (j) m and φ (j) for each tree and node.…”

Section: The End-to-end Modelmentioning

confidence: 99%

“…The existing well-known statistical models (such as autoregressive integrated moving average and exponential smoothing) for regression are robust to overfitting, have generally fewer parameters to estimate, easier to interpret due to the intuitive nature of the underlying model and amenable to automated procedures for hyperparameter selection [7]. However, these models have very strong assumptions about the data such as stationarity and linearity, which prevent them from capturing nonlinear patterns that real life data tend to possess [8].…”

Section: Introductionmentioning

confidence: 99%

A Hybrid Framework for Sequential Data Prediction with End-to-End Optimization

Aydın¹,

Kozat²

2022

Preprint

View full text Add to dashboard Cite

We investigate nonlinear prediction in an online setting and introduce a hybrid model that effectively mitigates, via an end-to-end architecture, the need for hand-designed features and manual model selection issues of conventional nonlinear prediction/regression methods. In particular, we use recursive structures to extract features from sequential signals, while preserving the state information, i.e., the history, and boosted decision trees to produce the final output. The connection is in an end-to-end fashion and we jointly optimize the whole architecture using stochastic gradient descent, for which we also provide the backward pass update equations. In particular, we employ a recurrent neural network (LSTM) for adaptive feature extraction from sequential data and a gradient boosting machinery (soft GBDT) for effective supervised regression. Our framework is generic so that one can use other deep learning architectures for feature extraction (such as RNNs and GRUs) and machine learning algorithms for decision making as long as they are differentiable. We demonstrate the learning behavior of our algorithm on synthetic data and the significant performance improvements over the conventional methods over various real life datasets. Furthermore, we openly share the source code of the proposed method to facilitate further research.

show abstract

“…With the recent growth in machine learning research, forecasting methods and tools also improved, with more developments expected to come in the future [16]. Amershi et al suggest focusing not only on debugging but also on error analysis [1].…”

Section: Testing Machine Learningmentioning

confidence: 99%

TimeGym: Debugging for Time Series Modeling in Python

Seca¹

2021

Preprint

View full text Add to dashboard Cite

We introduce the TimeGym Forecasting Debugging Toolkit, a Python library for testing and debugging time series forecasting pipelines. Tim-eGym simplifies the testing forecasting pipeline by providing generic tests for forecasting pipelines fresh out of the box. These tests are based on common modeling challenges of time series. Our library enables forecasters to apply a Test-Driven Development approach to forecast modeling, using specified oracles to generate artificial data with noise.

show abstract

Forecasting: theory and practice

Cited by 397 publications

References 1,547 publications

Forecasting Electricity Prices

Forecasting Electricity Prices

A Hybrid Framework for Sequential Data Prediction with End-to-End Optimization

TimeGym: Debugging for Time Series Modeling in Python

Contact Info

Product

Resources

About