Forecasting, Volume 4, Issue 4 (December 2022) – 17 articles

In times of rapid change and rising human pressures on marine systems, information about the future state of the ocean can provide decision-makers with time to avoid adverse impacts and maximise opportunities. An ecological forecast predicts changes in ecosystems and its components due to environmental forcing such as climate variability and change, extreme weather conditions, pollution, or habitat change. Here, we summarise examples from several sectors and a range of locations. We describe the need, approach, forecast performance, delivery system, and end user uptake. This examination shows that near-term ecological forecasts are needed by end users, decisions are being made based on forecasts, and there is an urgent need to develop operational information systems to support sustainable ocean management. An operational information system is critical for connecting to decision makers and providing an enduring approach to forecasting and proactive decision making. These operational systems require significant investment and ongoing maintenance but are key to delivering ecological forecasts for societal benefits. Iterative forecasting practices could provide continuous improvement by incorporating evaluation and feedback to overcome the limitations of the imperfect model and incomplete observations to achieve better forecast outcomes and accuracy. Full article

The study investigated the empirical role of past values of Somalia’s GDP growth rates in its future realizations. Using the Box–Jenkins modeling method, the study utilized 250 in-sample quarterly time series data to forecast out-of-the-sample Somali GDP growth rates for fourteen quarters. Balancing between parsimony and fitness criteria of model selection, the study found Autoregressive Integrated Moving Average ARIMA (5,1,2) to be the most appropriate model to estimate and forecast the trajectory of Somali economic growth. The study sourced the GDP growth data from World Bank World Development Indicators (WDI) for the period between 1960 to 2022. The study results predict that Somalia’s GDP will, on average, experience 4 percent quarterly growth rates for the coming three and half years. To solidify the validity of the forecasting results, the study conducted several ARIMA and rolling window diagnostic tests. The model errors proved to be white noise, the moving average (MA) and Autoregressive (AR) components are covariances stationary, and the rolling window test shows model stability within a 95% confidence interval. These optimistic economic growth forecasts represent a policy dividend for the government of Somalia after almost a decade-long stick-and-carrot economic policies between strict IMF fiscal disciplinary measures and World Bank development investments on target projects. The study, however, acknowledges that the developments of current severe droughts, locust infestations, COVID-19 pandemic, internal political, and security stability, and that the active involvement of international development partners will play a crucial role in the realization of these promising growth projections. Full article

Aiming for the forecasting and predictability of their future development, corporations have developed appropriate strategies as a result of the necessity to optimize the distribution networks of new IT products over time. The necessity of diversifying manufacturing was brought on by the fierce competition between businesses and the sophisticated consumer demand trends for personalized items. For businesses looking to create more effective distribution networks for their products, mass adaptability may be advantageous. Fuzzy cognitive mapping (FCM), associations developed from web analytics data, and simulation results based on dynamic and agent-based simulation models work together to practically aid digital marketing experts, decision-makers and analysts in offering answers to their corresponding problems. In order to apply the measures in agent-based modeling, the current work is based on the gathering of web analysis data over a predetermined time period, as well as on identifying the influence correlations between measurements. Full article

Proactive coping involves efforts to prepare for future stressors and may have implications for physical responses to stress. We examined age differences in physical reactivity to daily stressors moderated by proactive coping in a coordinated analysis across two separate daily diary studies. Study 1 included data from 116 older (age range 60–90) and 107 younger (age range 18–36) adults on daily stressors and physical health symptoms for 8 consecutive days. Study 2 included data from 140 adults (age range 19–86) on daily stressors and self-rated physical health for 29 consecutive days. Participants in both studies reported on their proactive coping on the first day of the study. Physical reactivity was indexed via lagged multilevel models as increases in daily physical symptoms in Study 1 and decreases in daily physical health in Study 2 with corresponding increases in daily stressors. Results indicated that in both studies, younger adults with low proactive coping were more physically reactive to daily stressors compared to younger adults with high proactive coping. Proactive coping was associated with reduced physical reactivity to daily stressors among younger adults, consistent with the characterization of a high degree of control and ample opportunities at earlier phases of adulthood which are critical for accumulating resources to proactively cope. Full article

We investigate whether Lasso-type linear methods are able to improve the predictive accuracy of OLS in selecting relevant firm characteristics for forecasting the future cross-section of stock returns. Through extensive Monte Carlo simulations, we show that Lasso-type predictions are superior to OLS when type II errors are a concern. The results change if the aim is to minimize type I errors. Finally, we analyze the predictive performance of the competing methods on the US cross-section of stock returns between 1974 and 2020 and show that only small and micro-cap stocks are highly predictable throughout the entire sample. Full article

Intuitively, one would expect a more skillful forecast if predicting weather averaged over one week instead of the weather averaged over one day, and similarly for different spatial averaging areas. However, there are few systematic studies of averaging and forecast skill with modern forecasts, and it is therefore not clear how much improvement in forecast performance is produced via averaging. Here we present a direct investigation of averaging effects, based on data from operational numerical weather forecasts. Data is analyzed for precipitation and surface temperature, for lead times of roughly 1 to 7 days, and for time- and space-averaging diameters of 1 to 7 days and 100 to 4500 km, respectively. For different geographic locations, the effects of time- or space-averaging can be different, and while no clear geographical pattern is seen for precipitation, a clear spatial pattern is seen for temperature. For temperature, in general, time averaging is most effective near coastlines, also effective over land, and least effective over oceans. Based on all locations globally, time averaging was less effective than one might expect. To help understand why time averaging may sometimes be minimally effective, a stochastic model is analyzed as a synthetic weather time series, and analytical formulas are presented for the decorrelation time. In effect, while time averaging creates a time series that is visually smoother, it does not necessarily cause a substantial increase in the predictability of the time series. Full article

In the smart grid (SG), user consumption data are increasing very rapidly. Some users consume electricity legally, while others steal it. Electricity theft causes significant damage to power grids, affects power supply efficiency, and reduces utility revenues. This study helps utilities reduce the problems of electricity theft, inefficient electricity monitoring, and abnormal electricity consumption in smart grids. To this end, an electricity theft dataset from the state grid corporation of China (SGCC) is employed and this study develops a novel model, a mixture of convolutional neural network and gated recurrent unit (CNN-GRU), for automatic power theft detection. Moreover, the hyperparameters of the proposed model are tuned using a meta-heuristic method, the cuckoo search (CS) algorithm. The class imbalance problem is solved using the synthetic minority oversampling technique (SMOTE). The clean data are trained and then tested with the proposed classification. Extensive simulations are performed based on real energy consumption data. The simulated results show that the proposed theft detection model (CNN-GRU-CS) solved the theft classification problem better than other approaches in terms of effectiveness and accuracy by 10% on average. The calculated accuracy of the proposed method is 92% and the precision is 94%. Full article

Credit scoring models help lenders decide whether to grant or reject credit to applicants. This paper proposes a credit scoring model based on boosted decision trees, a powerful learning technique that aggregates several decision trees to form a classifier given by a weighted majority vote of classifications predicted by individual decision trees. The performance of boosted decision trees is evaluated using two publicly available credit card application datasets. The prediction accuracy of boosted decision trees is benchmarked against two alternative machine learning techniques: the multilayer perceptron and support vector machines. The results show that boosted decision trees are a competitive technique for implementing credit scoring models. Full article

Forecasting daily and weekly passenger demand is a key fundamental process used by existing urban rail transit (URT) station authorities to diagnose operational problems and make decisions about train schedule patterns to improve operational efficiency, increase revenue management, and improve driving safety. The accuracy of the forecast results will directly affect the operation planning of urban rail transit (URT). Therefore, based on the collected inbound historical passenger data, this study used the Box–Jenkins time series with the Facebook Prophet algorithm to analyze the characteristics of urban rail transit passenger demand and achieved better computational forecasting performance accuracy. After analyzing the periodicity, correlation, and stationarity, different time series models were constructed. The Akaike information criteria (AIC), Bayesian information criteria (BIC), mean squared error (MSE), and root mean squared error (RMSE) were used to evaluate the adequacy of the best forecast model from among several tested candidates’ models for the Box–Jenkins. The parameters of the daily and weekly models were estimated using statistical software. The experimental results of this study are of both theoretical and practical significance to the urban rail transit (URT) station authorities for an effective station planning system. The forecasting results signify that the SARIMA (5, 1, 3) (1, 0, 0)²⁴ model performs better and is more stable in forecasting the daily passenger demand, and the ARMA (2, 1) model performs better in forecasting the weekly passenger demand. When comparing the SARIMA and ARMA models with the Facebook Prophet, results show that the Facebook Prophet model is superior to the SARIMA model for the daily time series, and the ARMA model is superior to the Facebook Prophet model for the weekly time series. Full article

The selection of an accurate performance metric is highly important to evaluate the quality of a forecasting method. This evaluation may help to select between different forecasting tools of forecasting outputs, and then support many decisions within a company. This paper proposes to evaluate the sensitivity and reliability of forecasts performance metrics. The methodology is tested using multiple time series of different scales and demand patterns, such as intermittent demand. The idea is to add to each series a noise following a known distribution to represent forecasting models of a known error distribution. Varying the parameters of the distribution of the noise allows to evaluate how sensitive and reliable performance metrics are to changes in bias and variance of the error of a forecasting model. The experiments concluded that sRMSE is more reliable than MASE in most cases on those series. sRMSE is especially reliable for detecting changes in the variance of a model and sPIS is the most sensitive metric to the bias of a model. sAPIS is sensible to both variance and bias but is less reliable. Full article

This paper provides new insights into the causal effects of the enlargement of the European Union (EU) on patent performance. The study focuses on the new EU member states (EU-13) and accession is considered as an intervention whose causal effect is estimated by the causal impact method using a Bayesian structural time-series model (proposed by Google). The empirical results based on data collected from the OECD database from 1985–2017 point towards a conclusion that joining the EU has had a significant impact on patent performance in Romania, Estonia, Poland, the Czech Republic, Croatia and Lithuania, although in the latter two countries, the impact was negative. For the rest of the EU-13 countries, there is no significant effect on patent performance. Whether the EU accession effect is significant or not, the EU-13 are far behind the EU-15 (countries which entered the EU before 2004) in terms of patent performance. The majority of patents (98.66%) are assigned to the EU-15, with just 1.34% of assignees belonging to the EU-13. Full article

Rainfall forecasting plays a key role in mitigating environmental risks in urban areas, which are subject to increasing hydrogeological risk due to transformations in the urban landscape. We present a new technique for probabilistic precipitation nowcasting by generating an ensemble of equiprobable forecasts, which is especially useful for weather radars with limited spatial range, and that can be used operationally on devices with low computational capacity. The ensemble members are obtained by a novel stochastic noise generation process, consistent with the spatial scales not resolved by the prediction model, which allows continuous downscaling of the output of a deep generative neural network. Through an in-depth analysis of the results, for precipitation accumulated in the first hour, measured by all the most robust skill indicators, extended to an entire year of data at 5-min time resolution, we demonstrate that the proposed procedure is able to provide calibrated, reliable, and sharp ensemble rainfall forecasts with a quality comparable or superior to the state-of-the-art classical alternative optical flow technique. The ensemble generation procedure we propose is sufficiently general to be applied in principle to other deterministic architectures as well, thus enabling their generalization in probabilistic terms. Full article

The main purpose of coherent mortality models is to produce plausible, joint forecasts for related populations avoiding, e.g., crossing or diverging mortality trajectories; however, the coherence assumption is very restrictive and it enforces trends that may be at odds with data. In this paper we focus on coherent, two-sex mortality models and we prove, under suitable conditions, that the coherence assumption implies sex gap unimodality, i.e., we prove that the difference in life expectancy between women and men will first increase and then decrease. Moreover, we demonstrate that, in the model, the sex gap typically peaks when female life expectancy is between 30 to 50 years. This explains why coherent mortality models predict narrowing sex gaps for essentially all Western European countries and all jump-off years since the 1950s, despite the fact that the actual sex gap was widening until the 1980s. In light of these findings, we discuss the current role of coherence as the gold standard for multi-population mortality models. Full article

The sudden onset of the COVID-19 pandemic posed significant challenges for forecasting professionals worldwide. This article examines the early forecasts of COVID-19 transmission, using the context of the United States, one of the early epicenters of the crisis. The article compares the relative accuracy of selected models from two forecasters who informed government policy in the first three months of the pandemic, the Institute of Health Metrics and Evaluation (IHME) and Columbia University. Furthermore, we examine whether the forecasts improved as more data became available in the subsequent months of the pandemic, using the forecasts from Los Alamos National Laboratory and the University of Texas, Austin. The analysis focuses on mortality estimates and compares forecasts using epidemiological and curve-fitting models during the first wave of the pandemic from March 2020 to October 2020. As health agencies worldwide struggled with uncertainty in models and projections of COVID-19 caseload and mortality, this article provides important insights that can be useful for crafting policy responses to the ongoing pandemic and future outbreaks. Full article

Artificial-intelligence-based algorithms are used in manufacturing to automate difficult activities and discover workflow or process patterns that had never been noticed before. Recent studies deal with the forecasting of the fracture location in dissimilar friction-stir-welded AA5754–C11000 alloys. Four types of supervised machine-learning-based classification algorithms i.e., decision tree, logistic classification, random forest, and AdaBoost were implemented. Additionally, in the present work, for the first time, a neurobiological-based unsupervised machine learning algorithm, i.e., self-organizing map (SOM) neural network, is implemented for determining the fracture location in dissimilar friction-stir-welded AA5754–C11000 alloys. Tool shoulder diameter (mm), tool rotational speed (RPM), and tool traverse speed (mm/min) are input parameters, while the fracture location, i.e., whether the specimen’s fracture is in the thermo-mechanically affected zone (TMAZ) of copper, or if it fractures in the TMAZ of aluminium. The results show that out of all implemented algorithms, the SOM algorithm is able to predict the fracture location with the highest accuracy of 96.92%. Full article

Big data technology and predictive analytics exhibit advanced potential for business intelligence (BI), especially for decision-making. This study aimed to explore current research studies, historic developing trends, and the future direction. A bibliographic study based on CiteSpace is implemented in this paper, 681 non-duplicate publications are retrieved from databases of Web of Science Core Collection (WoSCC) and Scopus from 2000 to 2021. The countries, institutions, cited authors, cited journals, and cited references with the most academic contributions were identified. Social networks and collaborations between countries, institutions, and scholars are explored. The cross degree of disciplinaries is measured. The hotspot distribution and burst keyword historic trend are explored, where research methods, BI-based applications, and challenges are separately discussed. Reasons for hotspots bursting in 2021 are explored. Finally, the research direction is predicted, and the advice is delivered to future researchers. Findings show that big data and AI-based methods for BI are one of the most popular research topics in the next few years, especially when it applies to topics of COVID-19, healthcare, hospitality, and 5G. Thus, this study contributes reference value for future research, especially for direct selection and method application. Full article

This study aims to forecast extreme fluctuations of Bitcoin returns. Bitcoin is the first decentralized and the largest, in terms of capitalization, cryptocurrency. A well-timed and precise forecast of extreme changes in Bitcoin returns is key to market participants since they may trigger large-scale selling or buying strategies that may crucially impact the cryptocurrency markets. We term the instances of extreme Bitcoin movement as ‘spikes’. In this paper, spikes are defined as the returns instances that outreach a two-standard deviations band around the mean value. Instead of the unconditional historic standard deviation that is usually used, in this paper, we utilized a GARCH(p,q) model to derive the conditional standard deviation. We claim that the conditional standard deviation is a more suitable measure of on-the-spot risk than the overall standard deviation. The forecasting operation was performed using the support vector machines (SVM) methodology from machine learning. The most accurate forecasting model that we created reached 79.17% out-of-sample forecasting accuracy regarding the spikes cases and 87.43% regarding the non-spikes ones. Full article