Treating and Pruning: New approaches to forecasting model selection and combination using prediction intervals

Meira, Erick; Oliveira, Fernando Luiz Cyrino; Jeon, Junho

doi:10.1016/j.ijforecast.2020.07.005

Cited by 27 publications

(12 citation statements)

References 34 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This conclusion is confirmed in [4][5][6][7] and in this paper. But narrowing of prediction intervals for combination of models is still investigated [12,13] and is in the scope of further research.…”

Section: Discussionmentioning

confidence: 99%

“…Implementation of non-linear combination forecasters [12] is an idea of next stage of research in this area comparing to linear combination technique used in [4][5][6][7]. Competitive selection of best models by their prediction intervals quality is under investigation of [13]. So, this problem is in scope of modern research.…”

Section: Time Series Forecast Errors and Prediction Intervalsmentioning

confidence: 99%

See 1 more Smart Citation

Improvement of time series forecasting quality by means of multiple models prediction averaging

2021

Proceedings of the III International Workshop on Modeling, Information Processing and Computing (MIP: Computing-2021)

View full text Add to dashboard Cite

Construction of time series models is usually based on Akaike and Bayes information criteria. Requirement of model simplicity is built inside the information criteria structure and the best fitted models aren't always best in terms of criteria values. Often there are a few best models that fit investigated time series well and there's problem of choice between them. Usually information criteria values allow to choose among them. But if one needs the best model by forecasts quality, best fitted models, or there are other thoughts a research has to choose and test models manually. At the same time when a few models are chosen it's possible to construct their combination. The simplest way is to count mean value of their forecasts and to use it as a combined prediction. Practical researches confirm that forecast error gets lower in this approach. Also, more complex construction than averaging of forecasts can be used (for example, weighted voting that is widely used in bagging technique in solution of classification problems). But this approach hasn't got enough theoretical base. From theoretical point of view confidence intervals of time series forecasts (usually here they're considered as prediction intervals) is also very complex task. Intervals tend to be very wide even for models with good prediction quality in terms of mean forecast errors. Thus, prediction intervals are rare in use. In this paper a few time series models are constructed for wage and income indices of Russian macroeconomic time series. Their predictions are combined into one forecast and it's quality is compared to individual ones. Transformation of forecast variance and prediction intervals in case of simple (moving averages MA(q) and autoregressions AR(p) of low order) models is also considered but is a part of further work.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Time Series Forecast Errors and Prediction Intervalsmentioning

confidence: 99%

Improvement of time series forecasting quality by means of multiple models prediction averaging

2021

Proceedings of the III International Workshop on Modeling, Information Processing and Computing (MIP: Computing-2021)

View full text Add to dashboard Cite

show abstract

“…Usually time series models are built in two stages: there's a training set used to construct them and a test set to evaluate their quality. But in case of bagging one needs three parts or training set should be subdivided in some parts to use them in construction of the models and in evaluation of their combinations [9][10][11]. In this research combination of forecasts is investigated from the prediction intervals point of view.…”

Section: Introductionmentioning

confidence: 99%

“…In this research combination of forecasts is investigated from the prediction intervals point of view. Behaviour of prediction intervals is often out of scope of research in a lot of ensemble models but this question is especially important in case of time series forecasting [11][12]. In the paper ARIMA models are considered.…”

Section: Introductionmentioning

confidence: 99%

Modeling of forecasts variance reduction at multiple time series prediction averaging with ARMA (1, q) functions

Musatov¹,

Petrusevich²

2022

Proceedings of the v International Workshop on Modeling, Information Processing and Computing (MIP: Computing-v 2022)

View full text Add to dashboard Cite

Combination of time series forecasts is usually considered as good technique in practice. But it has got weak theoretical explanation. In this research variance of time series forecasts and variance of combined models are considered. One is interested in the view of variance of forecasts function over one, two and three periods. Conditions which can lead to improvement of averaged time series predictions are in scope of this research. In this paper a few examples of the most popular time series models are observed: the moving average models MA(q), the autoregressions AR(p) models and their combination in the form of ARIMA(p, d, q) or ARMA(p, q) model. In particular, AR(1) and ARMA(1, q) are investigated. Nowadays there are researches about time series averaging. Approaches based on bagging and boosting are implemented very often in classification and regressions. It’s very appealing to use such strategy in tine series modeling. At the same time it’s easy to construct learning set and test set in classification tasks. But it’s a complex task in case of time series processing. There’s a need of two sets: to train time series models and to construct their combinations. Thus nowadays combination of time series models, combination of their forecasts or of their prediction intervals are in scope of view of a few complex researches. In this paper we investigate behaviour of time series predictions’ variance in order to have another useful approach in time series prediction averaging. Russian macroeconomical time series statistics is used as experimental time series.

show abstract

“…In the same paper, Grushka-Cockayne and Jose investigated six simple heuristics of combining intervals (and reference many more) which improve the situation, but they offered no theoretical explanation. Meira et al (2021) explored the area indepth and concluded that "no work has considered looking at the information delivered by prediction intervals when conducting model selection and/or combination. "…”

Section: Introductionmentioning

confidence: 99%

On the Uncertainty of a Combined Forecast: The Critical Role of Correlation

Magnus

Vasnev

2021

SSRN Journal

View full text Add to dashboard Cite

The purpose of this paper is to show that the effect of the zero-correlation assumption in combining forecasts can be huge, and that ignoring (positive) correlation can lead to confidence bands around the forecast combination that are much too narrow. In the typical case where three or more forecasts are combined, the estimated variance increases without bound when correlation increases. Intuitively, this is because similar forecasts provide little information if we know that they are highly correlated. Although we concentrate on forecast combinations and confidence bands, our theory applies to any statistic where the observations are linearly combined. We apply our theoretical results to explain why forecasts by Central Banks (in our case, the Bank of Japan) are so frequently misleadingly precise. In most cases, a correlation above 0.7 is required to produce reasonable confidence bands.

show abstract

Treating and Pruning: New approaches to forecasting model selection and combination using prediction intervals

Cited by 27 publications

References 34 publications

Improvement of time series forecasting quality by means of multiple models prediction averaging

Improvement of time series forecasting quality by means of multiple models prediction averaging

Modeling of forecasts variance reduction at multiple time series prediction averaging with ARMA (1, q) functions

On the Uncertainty of a Combined Forecast: The Critical Role of Correlation

Contact Info

Product

Resources

About