Forecast Window Research Articles

Comparing statistical forecasting methods for modelling respiratory virus healthcare demand

Abstract Background Emergency departments in the United Kingdom are under sustained pressure with rising demand for healthcare, gradually decreasing the functionality of the NHS. Forecasts can aid management in the allocation of appropriate healthcare supplies and staffing needs in advance of an expected surge, as well as allow stakeholders to assess most likely outcomes and potential worst-case scenarios. This is especially valuable during the seasonal waves of respiratory infections. Respiratory viruses are usually concentrated in a 6-8 week seasonal surge during the winter period, creating large pressure on the health system. Current forecast models’ performance depreciates significantly after 1 week. This analysis will forecast 6 weeks of respiratory health admissions, comparing traditional and newly emerging forecasting methods. Methods Using 20 years of English hospital admissions data, we evaluated both within season training models (2-20 weeks) and longer term seasonal models (1-10 years), to forecast 6 weeks of hospital admissions. Respiratory admissions were defined according to ARI/ILI ICD10 codes assigned in hospital, stratified by age and region. Models were iteratively constructed, trained and probabilistic forecasts were evaluated across the respiratory season of September to April using Weighted Interval Score. Averages of each model across iterations were generated for comparison. Results For non-seasonal models, shorter model training lengths created better forecast models for 6 weeks. For seasonal models, optimal training length varied across model complexity. Simple models performed best at forecasting 1 week, however performance deteriorated the most as the forecast window increased. Overall, the Prophet model performed best on average at forecasting 6 weeks across all scenarios. Conclusions Applying our framework for optimal model selection in forecasting analysis will help to inform better practice for future health demand modelling. Key messages • This analysis will inform better practice for future health demand forecasting 6 weeks in advance, allowing for preparation of resources for seasonal surges in healthcare demand. • This analysis suggests using models with modest levels of increased complexity over simpler models could significantly improve forecast performance in the short-medium term.

RUL forecasting for wind turbine predictive maintenance based on deep learning

Predictive maintenance (PdM) is increasingly pursued to reduce wind farm operation and maintenance costs by accurately predicting the remaining useful life (RUL) and strategically scheduling maintenance. However, the remoteness of wind farms often renders current methodologies ineffective, as they fail to provide a sufficiently reliable advance time window for maintenance planning, limiting PdM's practicality. This study introduces a novel deep learning (DL) methodology for future-RUL forecasting. By employing a multi-parametric attention-based DL approach that bypasses feature engineering, thereby minimizing the risk of human error, two models—ForeNet-2d and ForeNet-3d—are proposed. These models successfully forecast the RUL for seven multifaceted wind turbine (WT) failures with a 2-week forecast window. The most precise forecast deviated by only 10 minutes from the actual RUL, while the least accurate prediction deviated by 1.8 days, with most predictions being off by only a few hours. This methodology offers a substantial time frame to access remote WTs and perform necessary maintenance, thereby enabling the practical implementation of PdM.

Observable or latent Markov chains for score-driven regime-switching volatility?

ABSTRACT We study the statistical and forecasting performances of two regime-switching Beta-t-EGARCH (exponential generalized autoregressive conditional heteroscedasticity) models, i.e. observable-switching (OS) Beta-t-EGARCH and Markov-switching (MS) Beta-t-EGARCH. Both are non-path-dependent score-driven regime-switching volatility models, and their regime-switching specifications can be related to corresponding non-path-dependent Markov-switching GARCH (MS-GARCH) specifications. We present the estimation procedures for OS-Beta-t-EGARCH and MS-Beta-t-EGARCH. We use data on the weekly log-returns of the Standard & Poor’s 500 (S&P 500) index and a random sample of 50 stocks from the S&P 500 from March 1986 to July 2024 ( T = 2 , 000 ). The out-of-sample forecasting window is from May 2005 to July 2024 ( T f = 1 , 000 ). We compare the in-sample statistical and out-of-sample density forecasting performances of Beta-t-EGARCH, OS-Beta-t-EGARCH, and MS-Beta-t-EGARCH. We find that the statistical and density forecasting performances of OS-Beta-t-EGARCH are superior to MS-Beta-t-EGARCH, motivating its practical use by investors and risk managers.

A statistic of the subseasonal forecast skill windows of 2-meter air temperature

The forecast skill on subseasonal time scale is currently limited, inhibiting its effective application broadly. Under some special conditions, the subseasonal-seasonal (S2S) forecast skill would be increased intermittently, namely, the forecast opportunity window. The identification of such forecast windows can increase the credibility of subseasonal prediction results, providing final users the opportunity to make the appropriate decision. To identify the subseasonal forecast window of 2-m air temperature (T2m), based on the S2S multi-model prediction results, this study evaluated the pattern correlation coefficient between the third-week T2m forecasted by S2S models and the observation. When the forecast skills of the majority of models reach the prescribed threshold, this period is defined as a forecast skill window. In this way, the subseasonal forecast skill windows of T2m over different regions of the global lands are identified. From the perspective of seasonal distribution, the forecast skill windows over almost all the continents appear more in boreal winter, while the forecast skill windows of the Australian continent mainly appear in boreal summer. Significant differences can be found in the occurrence frequency of forecast windows during different phases of ENSO. During El Niño events, forecast windows appear more frequently over North America, Asia, and South America, especially during winter-spring from January to April. From the T2m spatial pattern during the window periods, the forecast skill windows have some relevance among several continents, and the windows over the whole global land mainly correspond to those over Asia, Europe, and North America. Deep investigation of the physical mechanism behind the forecast skill windows helps understand the sources of predictability and improves the skill of subseasonal prediction.

Multisite Long-Term Photovoltaic Forecasting Model Based on VACI

In the field of photovoltaic (PV) power prediction, long-term forecasting, which is more challenging than short-term forecasting, can provide more comprehensive and forward-looking guidance. Currently, significant achievements have been made in the field of short-term forecasting for PV power, but inadequate attention has been paid to long-term forecasting. Additionally, multivariate global forecasting across multiple sites and the limited historical time series data available further increase the difficulty of prediction. To address these challenges, we propose a variable–adaptive channel-independent architecture (VACI) and design a deep tree-structured multi-scale gated component named DTM block for this architecture. Subsequently, we construct a specific forecasting model called DTMGNet. Unlike channel-independent modeling and channel-dependent modeling, the VACI integrates the advantages of both and emphasizes the diversity of training data and the model’s adaptability to different variables across channels. Finally, the effectiveness of the DTM block is empirically validated using the real-world solar energy benchmark dataset. And on this dataset, the multivariate long-term forecasting performance of DTMGNet achieved state-of-the-art (SOTA) levels, particularly making significant breakthroughs in the 720-step ultra-long forecasting window, where it reduced the MSE metric below 0.2 for the first time (from 0.215 to 0.199), representing a reduction of 7.44%.

Volatility forecasting on China's oil futures: New evidence from interpretable ensemble boosting trees

This paper investigates the performance of ensemble boosting trees in forecasting volatility of China's crude oil futures by combining rich feature variables and multiple volatility forecasting models. The empirical results demonstrate that ensemble boosting tree models significantly outperform the HAR-RV model and traditional machine learning models, with the CatBoost and the LightGBM having the best forecasting performance, and that these conclusions hold up under robustness tests. Using the SHAP values model interpretability instrument, this paper analyzes the model interpretability of LightGBM and CatBoost in terms of the drivers of volatility forecasting, the contribution of variables in a specific period, and the performance of variables in forecasting outliers. It is discovered that macroeconomic variables and HAR-type variables have different forecasting contributions in CatBoost and LightGBM, and that the contribution of different variables to the forecasting window varies significantly within a single interval. In addition, the paper concludes that there is heterogeneity in the forecast contribution of the same predictor across models, so the selection of variables for forecasting volatility should be based on the actual situation. Lastly, additional analysis confirms that the ensemble boosting tree models also have a high economic value.

Data-Driven Modeling and Analysis for COVID-19 Pandemic Hospital Beds Planning

The COVID-19 pandemic presents unprecedented challenges for the US healthcare system, and the critical care settings are heavily impacted by the pressures of caring for COVID-19 patients. However, hospital pandemic preparedness has been hampered by a lack of disease specific planning guidelines. In this paper, we proposed a holistic modeling and analysis approach, with a system dynamics model to predict COVID-19 cases and a discrete-event simulation to evaluate hospital bed utilization, to support the hospital planning decisions. Our model was trained using the public data from the JHU Coronavirus Resource Center and was validated using historical patient census data from the University of Florida Health Jacksonville, Jacksonville, FL and public data from the Florida Department of Health (FDOH). Various experiments were conducted to investigate different control measures and the variants of the virus and their impact on the disease transmission, and subsequently, the hospital planning needs. Our proposed approach can be tailored to a given hospital setting of interest and is also generalizable to other hospitals to tackle the pandemic planning challenge. <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Note to Practitioners</i> —We proposed a holistic modeling and analysis approach to support hospital preparedness and resource planning during the COVID-19 pandemic. To capture the highly dynamic pandemic environment, we developed a numerical method to estimate <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$R_0$</tex-math> </inline-formula> , the effective basic reproductive rate, and used the most recent estimated data series of daily <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$R_0$</tex-math> </inline-formula> to project the change in <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$R_0$</tex-math> </inline-formula> in a short-term forecast window. The prediction of the daily confirmed cases in that forecast window were then obtained based on recursively solving the system dynamics model, and was validated to be very close to the real confirmed cases from the public record. This data-driven approach allows us to gain a systematic understanding of the common trends across different states and regions, and to evaluate the effect of the control measures like the stay-at-home order and the impact of the virus variants on the disease transmission behavior. Furthermore, the dynamic prediction allows us to evaluate the hospital resource needs during different stages of the pandemic. The insights obtained through this effort shed light on the impact of interventions (e.g., vaccines and control measures) on the hospital preparedness to support appropriate hospital resource allocation.

Predictive neural representations of naturalistic dynamic input

Adaptive behavior such as social interaction requires our brain to predict unfolding external dynamics. While theories assume such dynamic prediction, empirical evidence is limited to static snapshots and indirect consequences of predictions. We present a dynamic extension to representational similarity analysis that uses temporally variable models to capture neural representations of unfolding events. We applied this approach to source-reconstructed magnetoencephalography (MEG) data of healthy human subjects and demonstrate both lagged and predictive neural representations of observed actions. Predictive representations exhibit a hierarchical pattern, such that high-level abstract stimulus features are predicted earlier in time, while low-level visual features are predicted closer in time to the actual sensory input. By quantifying the temporal forecast window of the brain, this approach allows investigating predictive processing of our dynamic world. It can be applied to other naturalistic stimuli (e.g., film, soundscapes, music, motor planning/execution, social interaction) and any biosignal with high temporal resolution.

Chinese agricultural futures volatility: New insights from potential domestic and global predictors

This paper investigates whether the potential predictors from China and globally can efficiently predict Chinese agricultural futures volatility by adopting the REGARCH-MIDAS framework. We highlight the predictability of numerous Chinese potential predictors for forecasting ten agricultural futures volatility, which is relatively better than that of global potential predictors. Robustness tests such as different realized measure and different forecasting window confirm the above conclusions. Performances of predictors during different volatility levels, before and during the COVID-19 pandemic are further discussed. This paper tries to shed new light on the volatility prediction of Chinese agricultural futures markets.

An assessment of stingless beehive climate impact using multivariate recurrent neural networks

&lt;p&gt;A healthy bee colony depends on various elements, including a stable habitat, a sufficient source of food, and favorable weather. This paper aims to assess the stingless beehive climate data and examine the precise short-term forecast model for hive weight output. The dataset was extracted from a single hive, for approximately 36-hours, at every seven seconds time stamp. The result represents the correlation analysis between all variables. The evaluation of root-mean-square error (RMSE), as well as the RMSE performance from various types of topologies, are tested on four different forecasting window sizes. The proposed forecast model considers seven of input vectors such as hive weight, an inside temperature, inside humidity, outside temperature, outside humidity, the dewpoint, and bee count. The various network architecture examined for minimal RMSE are long short-term memory (LSTM) and gated recurrent units (GRU). The LSTM1X50 topology was found to be the best fit while analyzing several forecasting windows sizes for the beehive weight forecast. The results obtained indicate a significant unusual symptom occurring in the stingless bee colonies, which allow beekeepers to make decisions with the main objective of improving the colony’s health and propagation.&lt;/p&gt;

Have the predictability of oil changed during the COVID-19 pandemic: Evidence from international stock markets

The COVID-19 has undoubtfully brought fierce shocks to the real economic activities, financial market and public lives. Under this special condition, this study explores whether the predictability of crude oil futures information has changed before and during the COVID-19 pandemic for 19 international stock markets. From an in-sample perspective, we find that the crude oil futures RV can significantly affect future stock volatility for each equity index except SSEC. Moreover, the out-of-sample results from statistic and economic perspective reveal that crude oil futures RV is a more efficient predictor during the COVID-19 pandemic compared with the pre-crisis period. Furthermore, we find that the predictability of crude oil futures information is stronger from March to May 2020, when the epidemic is seriously prevailing. The empirical results from alternative evaluation method, recursive window method, alternative realized measures, controlling VIX and the seasonal effect, asymmetric forecasting window and different testing windows are robust and consistent. Our findings could offer novel and significant policy and practical implications.

An Analysis of Storage Requirements and Benefits of Short-Term Forecasting for PV Ramp Rate Mitigation

As renewable energy penetration on the grid increases, requirements are being placed on PV owners and operators to limit power ramp rates. PV power ramping is an issue for grid stability because generation load balance must be continually met. When a large PV resource significantly increases or decreases, another resource must compensate. To aid in grid stability, ramp rate limitations have been imposed on PV plants. This article addresses how much fast-responding storage is needed to mitigate high ramp rates of PV plants, and how much benefit is there from short-term power forecasting in terms of reducing the storage requirement. The results provide a baseline estimate for system planners and designers. Furthermore, the storage controller design and optimization are given, along with the open-source code, such that others can tailor the simulation to their specific plant and weather profile. Results from studying a 100 MW PV plant power production profile show a reduction in ramp-rate violations from 10% of yearly intervals to below 1% with 12 min of storage. With forecasting, the same level of smoothing is achieved with a 5-min rated storage. A sensitivity analysis shows the impacts of varying constraints, such as storage power rating, PV system size for geographic smoothing, forecast window length, and the ramp-rate limit magnitude.

Forecasting of Continuous Vital Sign Using Multivariate Auto-Regressive Models.

This project assessed the use of multivariate auto-regressive (MAR) models to create forecasts of continuous vital signs in hospitalized patients. A total of 20 hours continuous (1/60Hz) heart rate and respiration rate from eight postoperative patients, where used to fit a centered MAR model for forecasting in windows of 15 minutes. The model was fitted using Markov Chain Monte Carlo sampling, and the model was evaluated on data from five additional patients. The results demonstrate an average RMSE in the forecast window of 11.4 (SD: 7.30) beats per minute for heart rate and 3.3 (SD:1.3) breaths per minute for respiration rate. These results indicate potential for forecasting vital signs in a clinical setting.

EcoForecast: An interpretable data-driven approach for short-term macroeconomic forecasting using N-BEATS neural network

It will be beneficial to devise an effective approach for short-term macroeconomic forecasting. Existing traditional statistics-based macroeconomic forecasting mainly focuses on exploring feasible methods for improving the accuracy of long-term predictions. However, the performance of short-term predictions was far less impressive under the impact of unexpected incidents. Furthermore, some deep learning-based approaches can achieve fine-grained variable frequency forecasting and preliminarily demonstrate effective results, but the interpretability is still controversial. Therefore, how to consider both the performance and the interpretability has already become a universal concern and an urgent unsolved problem. In this paper, we identified the above issue and proposed an interpretable data-driven approach, named EcoForecast11The code and data are available at https://github.com/navfour/ecoforecast., for short-term macroeconomic forecasting based on the N-BEATS (neural basis expansion analysis for interpretable time series forecasting) neural network. To the best of our knowledge, EcoForecast is the first interpretable purely data-driven unified normative scheme for macroeconomic forecasting that achieves variable forecast frequencies and prediction domains, surpassing traditional statistics-based and deep learning-based approaches in performance or interpretability.EcoForecast used a three-level hierarchical signal encoding, including the fully connected neural network (FCNN) level, the block level, and the stack level. The FCNN level implemented both forecast and backcast information extraction for the temporal prediction and parameter learning of context. Block levels were connected by residuals so that the block’s backcast could be sequentially filtered on the input. The stack level was used to form the top-level system of EcoForecast, where each stack was constrained to specialize in different inductive functions. To some extent, EcoForecast can balance effectiveness, efficiency, generalizability and interpretability while conditions such as forecast frequency and time window change, even when unexpected incidents occur. Based on the actual macroeconomic data for China from 1992 to 2022, the data-driven EcoForecast demonstrated high stability in different sequence learning scenarios and the accompanying high-accuracy performance. This stability was reflected in smaller prediction error expectation and variance, tolerance of fewer input samples, and robustness across prediction domains. The experimental results indicated that EcoForecast improved the accuracy up to 3.94 times compared with the traditional BVAR. In the robustness test, EcoForecast required only a quarter of the data to achieve 2.51 times smaller forecast errors than the BVAR while also improving the accuracy of varied macroeconomic indicators such as the Purchasing Managers’ Index (PMI) and national electricity generation (ELEC)forecasting by 2.38 and 1.45 times. EcoForecast had a high sensitivity to the emergence of economic inflection points and adapts quickly when the economic environment changes, thus demonstrating a performance that exceeds traditional solutions in GDP forecasting during epidemics and PMI forecasting during economic turmoil. Interfacing with traditional economics research, the interpretable EcoForecast can uncover the trends and cycles of economic change, from which the conclusions are validated with the actual economics practice in China. Our findings can provide a new possible research direction for short-term macroeconomic forecasting.

Forecasting the Chinese low-carbon index volatility

This paper investigates the predictive power of economic policy uncertainty on the Chinese low-carbon market volatility and takes into account realized measures. First, in-sample analysis shows that both economic policy uncertainty and intraday high-frequency information have a significant impact on low-carbon index volatility. Second, out-of-sample evaluations show that the model combining China's economic policy uncertainty and intraday high-frequency information has the best predictive power. Finally, we use several robustness tests of alternative macroeconomic variable, alternative forecasting window, and alternative realized measure to prove that the results of this study are robust. This study enriches the market volatility model research. In addition, it can also promote low-carbon investment and provide a reference for national macro-control.

Decreasing False-alarm Rates in CNN-based Solar Flare Prediction Using SDO/HMI Data

A hybrid two-stage machine-learning architecture that addresses the problem of excessive false positives (false alarms) in solar flare prediction systems is investigated. The first stage is a convolutional neural network (CNN) model based on the VGG-16 architecture that extracts features from a temporal stack of consecutive Solar Dynamics Observatory Helioseismic and Magnetic Imager magnetogram images to produce a flaring probability. The probability of flaring is added to a feature vector derived from the magnetograms to train an extremely randomized trees (ERT) model in the second stage to produce a binary deterministic prediction (flare/no-flare) in a 12 hr forecast window. To tune the hyperparameters of the architecture, a new evaluation metric is introduced: the “scaled True Skill Statistic.” It specifically addresses the large discrepancy between the true positive rate and the false positive rate in the highly unbalanced solar flare event training data sets. Through hyperparameter tuning to maximize this new metric, our two-stage architecture drastically reduces false positives by ≈48% without significantly affecting the true positives (reduction by ≈12%), when compared with predictions from the first-stage CNN alone. This, in turn, improves various traditional binary classification metrics sensitive to false positives, such as the precision, F1, and the Heidke Skill Score. The end result is a more robust 12 hr flare prediction system that could be combined with current operational flare-forecasting methods. Additionally, using the ERT-based feature-ranking mechanism, we show that the CNN output probability is highly ranked in terms of flare prediction relevance.

Influence of trends on subseasonal temperature prediction skill

AbstractSubseasonal‐to‐seasonal (S2S) predictions have a wide range of applications. Improving forecasts on this time‐scale has therefore become a major effort. To evaluate their performance, these forecasts are routinely compared to a reference that forecasts the climatological distribution at any given time. This distribution is commonly assumed to be stationary over the verification period on time‐scales longer than the seasonal cycle. However, there are prominent deviations from this assumption, especially considering trends associated with climate change. By employing synthetic forecast‐verification pairs, we show that estimates of the probabilistic skill of both continuous and categorical forecasts increase as a function of the variance explained by the trend over the verification period, even when there are errors in the trends simulated by the forecasts. We also show this skill enhancement due to the trend in the ECMWF extended‐range ensemble prediction system. We demonstrate that the effects on the skill in an operational forecast setting are currently strongest in the Tropics. Our results show that care needs to be taken when evaluating forecasts that are subject to non‐stationarity on time‐scales much longer than the forecast verification window. This is especially important for determining the skill of categorical forecasts, where assumptions on the stationarity of the climatology enter both in the reference forecast and in the determination of the category thresholds. The results presented in this study are not exclusive to the S2S time‐scale but have wider implications for forecast verification on seasonal to decadal time‐scales, where the existence of trends can further impact forecast skill.

A Simple Statistical Intra-Seasonal Prediction Model for Sea Surface Variables Utilizing Satellite Remote Sensing

In this paper, a novel and simple statistical prediction model for sea surface multivariate is developed based on extended empirical orthogonal functions (referred to as the MEEOF model). This simple model embeds the temporal evolution information into the empirical orthogonal function spatial patterns, which effectively captures the spatial distribution of the sea surface variables and their evolution process over time, and can be used to improve the accuracy of intra-seasonal ocean forecasts. At the same time, it considers both the correlation between different spatial and temporal points and the dynamic balance between different sea surface variables. The performance of the MEEOF prediction model has been examined in the South China Sea (SCS) based on remote sensing satellite datasets. Experimental results demonstrate that this model has significant forecasting capability within the forecast window of 15–90 days, compared with the traditional persistence forecasts and optimal climatic normal (OCN) results. Furthermore, this model exhibits good forecast performance throughout the entire forecast window; the final prediction model (referred to as the fusion model) is established by obtaining the weighted average for the MEEOF forecasts and persistence forecast results. Numerical experimental results show that this fusion prediction model consistently outperforms the persistence model, the OCN model, and the linear regression model over the entire forecast window. A case study shows that the propagation of ocean waves and the coordination between different sea surface variables can be well predicted by this simple model, indicating that this fusion model has a potential advantage in intra-seasonal predictions of the ocean.

A Deep Learning Model for Forecasting Velocity Structures of the Loop Current System in the Gulf of Mexico

Despite the large efforts made by the ocean modeling community, such as the GODAE (Global Ocean Data Assimilation Experiment), which started in 1997 and was renamed as OceanPredict in 2019, the prediction of ocean currents has remained a challenge until the present day—particularly in ocean regions that are characterized by rapid changes in their circulation due to changes in atmospheric forcing or due to the release of available potential energy through the development of instabilities. Ocean numerical models’ useful forecast window is no longer than two days over a given area with the best initialization possible. Predictions quickly diverge from the observational field throughout the water and become unreliable, despite the fact that they can simulate the observed dynamics through other variables such as temperature, salinity and sea surface height. Numerical methods such as harmonic analysis are used to predict both short- and long-term tidal currents with significant accuracy. However, they are limited to the areas where the tide was measured. In this study, a new approach to ocean current prediction based on deep learning is proposed. This method is evaluated on the measured energetic currents of the Gulf of Mexico circulation dominated by the Loop Current (LC) at multiple spatial and temporal scales. The approach taken herein consists of dividing the velocity tensor into planes perpendicular to each of the three Cartesian coordinate system directions. A Long Short-Term Memory Recurrent Neural Network, which is best suited to handling long-term dependencies in the data, was thus used to predict the evolution of the velocity field in each plane, along each of the three directions. The predicted tensors, made of the planes perpendicular to each Cartesian direction, revealed that the model’s prediction skills were best for the flow field in the planes perpendicular to the direction of prediction. Furthermore, the fusion of all three predicted tensors significantly increased the overall skills of the flow prediction over the individual model’s predictions. The useful forecast period of this new model was greater than 4 days with a root mean square error less than 0.05 cm·s−1 and a correlation coefficient of 0.6.

Oil price volatility predictability: New evidence from a scaled PCA approach

We introduce the scaled principal component analysis (sPCA) method to forecast oil volatility, and compare it with two commonly used dimensionality reduction methods: principal component analysis (PCA) and partial least squares (PLS). By combining the simple autoregressive model with the three dimensionality reduction methods, we obtain several interesting and notable findings. First, the model with the sPCA diffusion index performs substantially better than the competing models based on the out-of-sample Roos2 test. Moreover, the model with the sPCA diffusion index consistently demonstrates superior forecasting power compared with the other models under different macroeconomic conditions (e.g., business cycle recessions and expansions, high- and low-volatility levels, and the COVID-19 pandemic). Furthermore, the findings of our study are strongly robust to various robustness tests, such as alternative forecasting window sizes and different lags of model selection.