Statistical Forecasting Models Research Articles

The Quagmire of Arrested Development in Tropical Cyclones

Abstract The 48-h intensity forecasts for Hurricane Pamela (2021) from numerical weather prediction models, statistical–dynamical aids, and forecasters were a major forecast bust with Pamela making landfall as a minor rather than major hurricane. From the satellite presentation, Pamela exhibited a symmetric pattern referred to as central cold cover (CCC) in the subjective Dvorak intensity technique. Per the technique, the CCC pattern is accompanied by arrested development in intensity despite the seemingly favorable convective signature. To understand forecast uncertainty during occurrences, central cold cover frequency from 2011 to 2021 is documented. From these cases, composites of longwave infrared brightness temperatures from geostationary satellites for CCC cases are presented, and the surrounding tropical cyclone large-scale environment is quantified and compared with other tropical cyclones at similar latitudes and intensities. These composites show that central cold cover has a consistent presentation, but varies in the preceding hours for storms that eventually intensify or weaken. And, the synoptic-scale environment surrounding the tropical cyclone thermodynamically supports the vigorous deep convection associated with CCC. Finally, intensity forecast errors from numerical weather prediction models and statistical–dynamical aids are examined in comparison to similar tropical cyclones. This work shows that guidance struggles during CCC cases with intensity errors from these models being in the lowest percentiles of performance, particularly for 24- and 36-h forecasts. Significance Statement The appearance of symmetric cold clouds near the center of developing tropical cyclones is most often associated with future intensification. This simple relationship is widely used by statistical tropical cyclone intensity forecast models. Here, we reexamine and confirm that one subjectively determined nighttime cold cyclone cloud pattern termed the “central cold cover” pattern in Vern Dvorak’s seminal technique for estimating tropical cyclone intensity from infrared satellite images is indeed related to slow or arrested development, and represents a failure mode for these simple forecast models.

Fire Incidents, Trends, and Risk Mitigation Framework of Electrical Vehicle Cars in Australia

Electric Vehicles (EVs) offer a promising solution to reduce the environmental impact compared to internal combustion engine vehicles. However, EV adoption in Australia has been hindered by concerns over fire safety. This study aims to comprehensively analyse EV fire risks and trends in Australia, including those related to charging stations and lithium-ion batteries. The research utilises secondary data from various reputable sources to develop statistical forecasting models, which estimate that Australia will have approximately 1.73 million EVs by 2030 and 15.8 million by 2050. The study reveals an average EV fire frequency of six fires per million EVs in Australia, aligning with the global average. Consequently, Australia is expected to experience 9 to 10 EV fire incidents annually in 2030, 37 to 42 EV fire incidents annually in 2040, and 84 to 95 EV fire incidents annually in 2050. To address these risks, an EV fire risk control framework is considered to identify and recommend appropriate measures for life safety, lithium-ion batteries, charging, EV handling, and EV locations. This research provides vital evidence for regulators, policymakers, and the fire industry to effectively manage EV fire risks and enhance preparedness for the growing EV market in Australia.

Effectiveness of generative learning strategies in independent study of english using mobile technologies

The aim of the research is to determine and evaluate the practical impact of the educational concept of generative learning in independent study of the English language with the development of the most appropriate and effective educational system. The following methods were used to achieve the aim of the research: EF SET testing, correlation analysis, statistical and mathematical forecast models. Empirical component of the research and the analytical interpretation of the results indicate a positive effect in the application of generative methods of learning English. It is quantified by an improved dynamics of learning English for students of the experimental group (relative to the indicators of the control group) —2.915 times. Performance for the final test for the group which learned using typical methods was 3.06%. The relevant indicators for the group that received focal knowledge by generative learning methods were 8.92%. The results of the study indicate the practical impact of the concept of generative learning on the effectiveness of learning English, and the significant potential for other applications. It is advisable to conduct repeated empirical studies in other educational institutions and other educational subjects taking into account the established facts. The results of those studies will be the ground for developing a generalized educational standard, which will be proposed by the Ministry of Education and Culture of Ukraine for integration and implementation into approved teaching methods.

Carbon Price Combination Forecasting Model Based on Lasso Regression and Optimal Integration

Accurate carbon price index prediction can delve deeply into the internal law of carbon price changes, provide helpful information to managers and decision makers, as well as improve the carbon market system. Nevertheless, existing methods for combination forecasting typically arbitrarily choose a certain set of single forecasting models. However, a particular selection of forecasting models do not apply to all data sets due to the nonlinearity and nonsmoothness of the carbon trading price series. Therefore, choosing suitable single forecasting models for the combination is crucial. Considering the limitations of the current study, this study constructs a combined carbon trading forecasting model based on Lasso regression and optimal integration. By invoking the Lasso regression model, we can select suitable single forecasting models for combination forecasting based on the variation patterns of different training sets. Meanwhile, ARIMA, NARNN, LSTM, and 11 other single forecasting models are screened in this study, including both traditional statistical forecasting models and artificial intelligence forecasting models. First, the carbon price index is predicted using 11 single prediction models. Furthermore, given the multi-collinearity of the single prediction series, this study employs Lasso regression to reduce the dimensions of the single prediction models, which are then used to construct an optimal combination prediction model. Finally, the proposed model is applied to SZA-2017 and SZA-2019 carbon price data in Shenzhen. The results demonstrate that the model developed in this study outperforms other benchmark prediction models in terms of prediction error and direction accuracy, showing the efficacy of the proposed method.

Prediction of ICU Patients' Deterioration Using Machine Learning Techniques.

Assessing vital sign measurements within hospital settings presents a valuable opportunity for data analysis and knowledge extraction. By generating adaptable, personalized prediction models of patient vital signs, these models can yield clinically relevant insights not achievable through population-based models. This study aims to compare several statistical forecasting models to determine their real-life applicability. The primary objectives of this paper are to evaluate whether the following measurements: blood pressure, oxygen saturation, temperature and heart rate can predict deterioration in Intensive Care Unit (ICU) patients. Additionally, we aim to identify which of these measurements contributes most significantly to our prediction. Lastly, we seek to determine the most accurate data mining technique for real-life data applications. This retrospective chart review study utilized data from patients admitted to the ICU at a tertiary hospital between January and December, 2019. Data mining techniques for prediction included logistic regression, support vector machine classifier, k-nearest neighbors (KNN), gradient boosting classifier, and Naive Bayes classifier. A comprehensive comparison of these techniques was performed, focusing on accuracy, precision, recall, and F-measure. To achieve the research objectives, the SelectKBest class was applied to extract the most contributory features for prediction. Blood pressure ranked first with a score of 9.98, followed by respiratory rate, temperature, and heart rate. Analysis of 653 patient records indicated that 129 patients expired, while 542 patients were discharged either to their homes or other facilities. Among the five training models, two demonstrated the highest accuracy in predicting patient deterioration or survival at 88.83% and 84.72%, respectively. The gradient boosting classifier accurately predicted 115 out of 129 expired patients, while the KNN correctly predicted 109 out of 129 expired patients. Machine learning has the potential to enhance clinical deterioration prediction compared to traditional methods. This allows healthcare professionals to implement preventative measures and improve patients' quality of life, ultimately increasing average life expectancy. Although our research focused exclusively on ICU patients, data mining techniques can be applied in various contexts both within and outside the hospital setting.

Application of a statistical model for predicting thermodynamic parameters for Laves hydride phases

The statistical forecast model developed to optimize the process of searching for materials that meet the conditions of specific tasks of metal hydride technologies has been tested for application to the Laves phase alloys. The chemical and phase compositions of initial alloys were tested with X-ray and EDX. Absorption/desorption isotherm were measured and the experimental values of thermodynamic parameters compared with results of model calculations.

Forecasting annual maximum water level for the Negro River at Manaus using dynamical seasonal predictions

Early and skilful prediction of the Negro River maximum water levels at Manaus is critical for effective mitigation measures to safeguard lives and livelihoods. Using dynamical seasonal prediction hindcasts, from six prediction centres, we investigate extending the lead time of previously developed statistical models, which issue forecasts in March for Manaus. The original statistical forecast models used observed rainfall as the major predictor. We advance the capability to issue skilful forecasts earlier, in February. We develop ensemble forecasts by combining predictor data from observations and seasonal hindcasts. We compare those forecasts against the original statistical forecast models and forecasts using the observed climatology or persistence of predictors. The ensemble-mean forecasts, issued in February, using European Centre for Medium-Range Weather Forecasts (ECMWF) hindcast input, perform similarly as the original forecasts issued in March and gain one month of lead time. The ECMWF-based ensemble forecasts skilfully predict the likelihood of water levels exceeding the severe flood level of 29 m. Forecast performance reduces and ensemble spread increases with increasing lead time from February to January. We conclude that forecasts for Manaus maximum water levels can be produced using combined input from observations and real-time ECMWF forecasts.

Сезонність та суїцид: роль регіональних особливостей динаміки суїцидів в Україні у 2005-2021 роках у розробці статистичної моделі їх прогнозування

Summary Background. Suicidal behavior has traditionally been considered by medicine to be a serious public health problem. Suicides are the 13th leading cause of death in the general population, claiming more than 58,000 lives in Europe each year. The National Institute of Mental Health (NIMH), the United Kingdom's Office for National Statistics, and other developed countries' statistics provide accurate observations of the dynamics of this indicator. State Statistics Service of Ukraine and Ptoukha Institute for Demography and Social Studies of the National Academy of Sciences of Ukraine only in 2021 provided complete information on the number of deaths and their causes in Ukraine for a long time - from 2005 to 2021, which made our study possible. Objective. The aim of our study was to establish and compare the features of regional dynamic characteristics of changes in the number of deaths due to intentional self-harm. Based on this goal, the tasks were: construction of time series of the such deaths number in the period 2005 - 2021 for each region of Ukraine; study and mathematical and statistical analysis of the dynamic characteristics of these time series; identification of possible causes that explain the identified features of the identified dynamics of the such deaths number; construction of a statistical forecast of the dynamics of the number of suicides for each region for one year with the possibility of its further verification. Methods and materials. Research material - data from the State Statistics Service of Ukraine and the Institute of Demography and Social Research on the number of deaths due to intentional self-harm during 2005 - 2021. Research methods - analysis of time series using autocorrelation analysis based on Leung-Box statistics and the method of seasonal exponential smoothing. Results. In order to study the trends in the dynamics of suicidal activity in the regions of Ukraine, an autocorrelation of absolute indicators of the number of completed suicides was carried out and correlograms of time series of suicide numbers for each region were constructed. Further analysis of the constructed correlograms indicated that they should be divided into four groups. According to the calculation of the coefficient of determination, it was found that a fairly high share of the total variation of the series, both Ukraine (average R2 = 0.656) and all individual regions (average R2 = 0.731 ± 0.051), can be explained by our model, and the model itself may evaluate as consistent. According to the results of the analysis, it is possible to say that for most regions of Ukraine and for the state as a whole, the periods with the highest mortality rates due to completed suicides are the period from March to May, and July, and, to a lesser extent, January 2022. The period characterized by a gradual decrease in the number of completed suicides - August - December, as well as a sudden decrease in the number of deaths in February and June. We refused to analyze the results of the forecast of separate regions, taking into account socio-demographic features, namely, a very high demographic mobility of a large part of the population of Ukraine. We assume that the unacceptable results of mathematical and statistical analysis of reports of mortality due to completed suicides in many areas may be caused by a distorted picture of reporting data due to socio-demographic reasons, so they do not fully reflect the true picture of phenomenon. Conclusions. The statistical forecasting model built by us assumes that the seasonal periodicity in September 2021 - August 2022 should be characterized by a decrease in the number of suicides during September - December 2021, February, June and August 2022 with an increase in their number in January, March - May and June 2022. The studied chronobiological aspects are promising for further research to improve general and regional suicide prevention programs in Ukraine. Keywords: suicides, seasonality, time series, exponential smoothing, autocorrelation analysis, statistical prognostic model

Forecasting the July Precipitation over the Middle-Lower Reaches of the Yangtze River with a Flexible Statistical Model

The multiple regression method is still an important tool for establishing precipitation forecast models with a lead time of one season. This study developed a flexible statistical forecast model for July precipitation over the middle-lower reaches of the Yangtze River (MLYR) based on the prophase winter sea surface temperature (SST). According to the characteristics of observed samples and related theoretical knowledge, some special treatments (i.e., more flexible and better-targeted methods) were introduced in the forecast model. These special treatments include a flexible MLYR domain definition, the extraction of indicative signals from the SST field, artificial samples, and the amplification of abnormal precipitation. Rolling forecast experiments show that the linear correlation between prediction and observation is around 0.5, more than half of the abnormal precipitation years can be successfully predicted, and there is no contradictory prediction of the abnormal years. These results indicate that the flexible statistical forecast model is valuable in real-life applications. Furthermore, sensitivity experiments show that forecast skills without these special treatments are obviously decreased. This suggests that forecast models can benefit from using statistical methods in a more flexible and better-targeted way.

Forecast of differential settlement of structures using data of monitoring system

The article analyzes the results of observations of uneven settlement of various buildings, obtained in the course of long-term observations with the help of automated systems of deformation monitoring. The main purpose of the work is to evaluate the possibilities of predicting sedimentary processes using statistical forecasting models. Emphasis is placed on approach using time series analysis algorithms. The results of simulation using the integrated autoregressive moving average model are demonstrated. To build a forecast, data on the foundation settlement over a 5-year period were used. Predictive values obtained for 1-year and 2-year intervals are compared with the data of real observations. The comparison showed that at the chosen forecasting horizon, the deviation of the predicted settlement value from the actually observed one is comparable to the accuracy of sediment measurement by hydraulic leveling sensors. The study showed the promise of using this approach to predict the integrity of structures within the framework of an automated system of deformation monitoring.

Forecasting Cotton Whitefly Population Using Deep Learning

Agricultural is the primary source on which a country's economy depends. Pakistan is the world's fourth-largest cotton producer, making it one of its top cash crops. They can't figure out what sort of cotton is best for their climate because of government regulations, crop risks, and a low literacy rate among the general public and farmers. Using a model that considers temperature, this system attempts to describe the different types of plants. More than anything else, the pest's population is concerned with its diversity, how the weather affects its population, and when it is high or low. The primary purpose of this study was to develop a framework that can handle the complex process of the population of cotton white fly. Our main aim was to know the population of insects with their eggs and with their parents. And another purpose is to find out information about the variety with a low population of insects. To get early knowledge about the people against the temperature. So, cotton yield could be increased, and fewer chemicals should be used. Therefore, farmers' income could be improved. So, we will make the best model for predicting whiteflies on cotton ARIMAX. This model's accuracy was nearly on par with the statistical forecasting models. Because ARIMAX is a statistical model, it may also be used for forecasting.

Demand forecasting: AI-based, statistical and hybrid models vs practice-based models - the case of SMEs and large enterprises

Demand forecasting is one of the biggest challenges of post-pandemic logistics. It appears that logistics management based on demand prediction can be a suitable alternative to the just-in-time concept. This study aims to identify the effectiveness of AI-based and statistical forecasting models versus practice-based models for SMEs and large enterprises in practice. The study compares the effectiveness of the practice-based Prophet model with the statistical forecasting models, models based on artificial intelligence, and hybrid models developed in the academic environment. Since most of the hybrid models, and the ones based on artificial intelligence, were developed within the last ten years, the study also answers the question of whether the new models have better accuracy than the older ones. The models are evaluated using a multicriteria approach with different weight settings for SMEs and large enterprises. The results show that the Prophet model has higher accuracy than the other models on most time series. At the same time, the Prophet model is slightly less computationally demanding than hybrid models and models based on artificial neural networks. On the other hand, the results of the multicriteria evaluation show that while statistical methods are more suitable for SMEs, the prophet forecasting method is very effective in the case of large enterprises with sufficient computing power and trained predictive analysts.

The National Hurricane Center Tropical Cyclone Model Guidance Suite

Abstract The National Hurricane Center (NHC) uses a variety of guidance models for its operational tropical cyclone track, intensity, and wind structure forecasts, and as baselines for the evaluation of forecast skill. A set of the simpler models, collectively known as the NHC guidance suite, is maintained by NHC. The models comprising the guidance suite are briefly described and evaluated, with details provided for those that have not been documented previously. Decay-SHIFOR is a modified version of the Statistical Hurricane Intensity Forecast (SHIFOR) model that includes decay over land; this modification improves the SHIFOR forecasts through about 96 h. T-CLIPER, a climatology and persistence model that predicts track and intensity using a trajectory approach, has error characteristics similar to those of CLIPER and D-SHIFOR but can be run to any forecast length. The Trajectory and Beta model (TAB), another trajectory track model, applies a gridpoint spatial filter to smooth winds from the National Centers for Environmental Prediction (NCEP) Global Forecast System (GFS) model. TAB model errors were 10%–15% lower than those of the Beta and Advection model (BAM), the model it replaced in 2017. Optimizing TAB’s vertical weights shows that the lower troposphere’s environmental flow provides a better match to observed tropical cyclone motion than does the upper troposphere’s, and that the optimal steering layer is shallower for higher-latitude and weaker tropical cyclones. The advantages and disadvantages of the D-SHIFOR, T-CLIPER, and TAB models relative to their earlier counterparts are discussed. Significance Statement This paper provides a comprehensive summary and evaluation of a set of simpler forecast models used as guidance for NHC’s operational tropical cyclone forecasts, and as baselines for the evaluation of forecast skill; these include newer techniques that extend forecasts to 7 days and beyond.

Regional Photovoltaic Power Forecasting Using Vector Autoregression Model in South Korea

Renewable energy forecasting is a key for efficient resource use in terms of power generation and safe grid control. In this study, we investigated a short-term statistical forecasting model with 1 to 3 h horizons using photovoltaic operation data from 215 power plants throughout South Korea. A vector autoregression (VAR) model-based regional photovoltaic power forecasting system is proposed for seven clusters of power plants in South Korea. This method showed better predictability than the autoregressive integrated moving average (ARIMA) model. The normalized root-mean-square errors of hourly photovoltaic generation predictions obtained from VAR (ARIMA) were 8.5–10.9% (9.8–13.0%) and 18.5–22.8% (21.3–26.3%) for 1 h and 3 h horizon, respectively, at 215 power plants. The coefficient of determination, R2 was higher for VAR, at 4–5%, than ARIMA. The VAR model had greater accuracy than ARIMA. This will be useful for economical and efficient grid management.

Modeling labor costs using artificial intelligence tools

PurposeConstruction projects in Malaysia are often delayed and over budget due to heavy reliance on labor. Linear regression (LR) models have been used in most labor cost (LC) studies, which are less accurate than machine learning (ML) tools. Construction management applications have increasingly used ML tools in recent years and have greatly impacted forecasting. The research aims to identify the most influential LC factors using statistical approaches, collect data and forecast LC models for improved forecasts of LC.Design/methodology/approachA thorough literature review was completed to identify LC factors. Experienced project managers were administered to rank the factors based on importance and relevance. Then, data were collected for the six highest ranked factors, and five ML models were created. Finally, five categorical indices were used to analyze and measure the effectiveness of models in determining the performance category.FindingsWorker age, construction skills, worker origin, worker training/education, type of work and worker experience were identified as the most influencing factors on LC. SVM provided the best in comparison to other models.Originality/valueThe findings support data-driven regulatory and practice improvements aimed at improving labor issues in Malaysia, with the possibility for replication in other countries facing comparable problems.

A Deterministic–Statistical Hybrid Forecast Model: The Future of the COVID-19 Contagious Process in Several Regions of Mexico

More than two years after the declaration of the COVID-19 pandemic, we are still experiencing contagious waves. As this is a long-lasting process, it becomes relevant to have a predictive tool to identify the intensively active places within a region. This study presents the development of a forecasting model applied to foresee the progress of the contagious process in Mexico and its regions. The method comprehends aspects of deterministic and probabilistic modeling. The deterministic part comprises the classical SIR model with some adjustments. The probabilistic part builds and populates a three-dimensional array, which is then used to describe and recall the probabilities of going from one status to another after some time, very much like a Markovian process. The process status is modeled as the combination of two conditions: the infection exponential growth parameter and a proxy variable we named “permissiveness” that accounts for all combined social activity factors affecting COVID-19 propagation. The results offer projections of the exponential growth parameter and the number of newly infected individuals for three weeks into the future. The proposed method’s capabilities allow for predicting newly COVID-19-infected individuals with reasonable precision while capturing the characteristic dynamics and behavior of the modeled system.

Hydrological Forecasting under Climate Variability Using Modeling and Earth Observations in the Naryn River Basin, Kyrgyzstan

The availability of water resources in Central Asia depends greatly on snow accumulation in the mountains of Tien-Shan and Pamir. It is important to precisely forecast water availability as it is shared by several countries and has a transboundary context. The impact of climate change in this region requires improving the quality of hydrological forecasts in the Naryn river basin. This is especially true for the growing season due to the unpredictable climate behavior. A real-time monitoring and forecasting system based on hydrological watershed models is widely used for forecast monitoring. The study’s main objective is to simulate hydrological forecasts for three different hydrological stations (Uch-Terek, Naryn, and Big-Naryn) located in the Naryn river basin, the main water formation area of the Syrdarya River. We used the MODSNOW model to generate statistical forecast models. The model simulates the hydrological cycle using standard meteorological data, discharge data, and remote sensing data based on the MODIS snow cover area. As for the forecast at the monthly scale, the model considers the snow cover conditions at separate elevation zones. The operation of a watershed model includes the effects of climate change on river dynamics, especially snowfall and its melting processes in different altitude zones of the Naryn river basin. The linear regression models were produced for monthly and yearly hydrological forecasts. The linear regression shows R2 values of 0.81, 0.75, and 0.77 (Uch-Terek, Naryn, and Big-Naryn, respectively). The correlation between discharge and snow cover at various elevation zones was used to examine the relationship between snow cover and the elevation of the study. The best correlation was in May, June, and July for the elevation ranging from 1000–1500 m in station Uch-Terek, and 1500–3500 m in stations Naryn and Big-Naryn. The best correlation was in June: 0.87; 0.76; 0.84, and May for the elevation ranging from 1000–3500 m in station Uch-Terek, and 2000–3000 m in stations Naryn and Big-Naryn. Hydrological forecast modeling in this study aims to provide helpful information to improve our under-standing that the snow cover is the central aspect of water accumulation.

LoMEF: A framework to produce local explanations for global model time series forecasts

Global forecasting models (GFMs) that are trained across a set of multiple time series have shown superior results in many forecasting competitions and real-world applications compared with univariate forecasting approaches. One aspect of the popularity of statistical forecasting models such as ETS and ARIMA is their relative simplicity and interpretability (in terms of relevant lags, trend, seasonality, and other attributes), while GFMs typically lack interpretability, especially relating to particular time series. This reduces the trust and confidence of stakeholders when making decisions based on the forecasts without being able to understand the predictions. To mitigate this problem, we propose a novel local model-agnostic interpretability approach to explain the forecasts from GFMs. We train simpler univariate surrogate models that are considered interpretable (e.g., ETS) on the predictions of the GFM on samples within a neighbourhood that we obtain through bootstrapping, or straightforwardly as the one-step-ahead global black-box model forecasts of the time series which needs to be explained. After, we evaluate the explanations for the forecasts of the global models in both qualitative and quantitative aspects such as accuracy, fidelity, stability, and comprehensibility, and are able to show the benefits of our approach.

A New Methodology to Produce More Skillful United States Cool-Season Precipitation Forecasts

Abstract The water resources of the western United States have enormous agricultural and municipal demands. At the same time, droughts like the one enveloping the West in the summer of 2021 have disrupted supply of this strained and precious resource. Historically, seasonal forecasts of cool-season (November–March) precipitation from dynamical models such as North American Multi-Model Ensemble (NMME) and the Seasonal Forecasting System 5 (SEAS5) from the European Centre for Medium-Range Weather Forecasts have lacked sufficient skill to aid in Western stakeholders’ and water managers’ decision-making. Here, we propose a new empirical–statistical framework to improve cool-season precipitation forecasts across the contiguous United States (CONUS). This newly developed framework is called the Statistical Climate Ensemble Forecast (SCEF) model. The SCEF framework applies a principal component regression model to predictors and predictands that have undergone dimensionality reduction, where the predictors are large-scale meteorological variables that have been prefiltered in space. The forecasts of the SCEF model captures 12.0% of the total CONUS-wide standardized observed variance over the period 1982/83–2019/20, whereas NMME captures 7.2%. Over the more recent period 2000/01–2019/20, the SCEF, NMME, and SEAS5 models respectively capture 11.8%, 4.0%, and 4.1% of the total CONUS-wide standardized observed variance. An important finding is that much of the improved skill in the SCEF, with respect to models such as NMME and SEAS5, can be attributed to better forecasts across most of the western United States.

An empirical study on probabilistic forecasting for predicting city-wide electricity consumption

In the competitive framework of the electricity markets, demand forecasting with high accuracy is invaluable for governmental institutions and power generation industry. By providing a basis for sound decision-making in energy management, it supports the decision makers to build capacity to meet the demand. In this research, we use electricity consumption data from 81 cities in Turkey to create global probabilistic forecasting models. For this purpose, we implement two prevailing probabilistic forecasting methods, namely, DeepAR and Deep State Space (DeepState) models. We compare the performances of these models against well-known statistical forecasting models (e.g., ARIMA), machine learning models (e.g., gradient boosting, and random forests), and deep learning models (e.g., long short term memory networks (LSTM) and gated recurrent unit networks (GRU)). Our results show that LSTM and GRU outperform others in terms of point forecasting performance, whereas DeepAR and DeepState models provide a better probabilistic forecasting performance for longer prediction horizons. We find that the global forecasting methods successfully leverage multiple time series (as in the case of 81-city forecasting data), however, mainly due to the relatively simple structure of the electricity datasets, they do not lead to meaningful performance gain over simpler models