Explore the vast range of titles available.

Weather forecasting is important for science and society. At present, the most accurate forecast system is the numerical weather prediction (NWP) method, which represents atmospheric states as discretized grids and numerically solves partial differential equations that describe the transition between those states 1 . However, this procedure is computationally expensive. Recently, artificial-intelligence-based methods 2 have shown potential in accelerating weather forecasting by orders of magnitude, but the forecast accuracy is still significantly lower than that of NWP methods. Here we introduce an artificial-intelligence-based method for accurate, medium-range global weather forecasting. We show that three-dimensional deep networks equipped with Earth-specific priors are effective at dealing with complex patterns in weather data, and that a hierarchical temporal aggregation strategy reduces accumulation errors in medium-range forecasting. Trained on 39 years of global data, our program, Pangu-Weather, obtains stronger deterministic forecast results on reanalysis data in all tested variables when compared with the world’s best NWP system, the operational integrated forecasting system of the European Centre for Medium-Range Weather Forecasts (ECMWF) 3 . Our method also works well with extreme weather forecasts and ensemble forecasts. When initialized with reanalysis data, the accuracy of tracking tropical cyclones is also higher than that of ECMWF-HRES. Three-dimensional deep neural networks can be trained to forecast global weather patterns, including extreme weather, with accuracy greater than or equal to that of the best numerical weather prediction models.

The purpose of this study is to examine the impact of environmental, social and governance (ESG) factors on stock liquidity and analyst forecast accuracy on the Johannesburg Stock Exchange (JSE) listed firms for the period 2012-2024. The empirical results reveal that market liquidity is positively associated with ESG and its components (environmental (ENV) and governance (GOV). Additionally, when investigating the relationship between ESG and analysts' forecast accuracy, we find that ENV and GOV scores improve forecast accuracy, while social (SOC) scores have a negative impact. Moreover, the results reveal that ESG disclosure serves as a critical strategic tool for maintaining financial stability by reducing BAS and ERROR during the pandemic crisis (COVID₁9). The study highlights the importance of ESG in enhancing liquidity and supporting sustainable finance, by providing significant implications for investors, analysts, and corporate stakeholders.

Purpose Property market models have the overriding aim of predicting reasonable estimates of key dependent variables (demand, supply, rent, yield, vacancy and net absorption rate). These can be based on independent drivers of core property and economic activities. Accurate predictions can only be conducted when ample quantitative data are available with fewer uncertainties. However, a broad-fronted social, technical and ecological evolution can throw up sudden, unexpected shocks that result in the econometric outputs sceptical to unknown risk factors. Therefore, the purpose of this paper is to evaluate Australian office market forecast accuracy and to determine whether the forecasts capture extreme downside risk events. Design/methodology/approach This study follows a quantitative research approach, using secondary data analysis to test the accuracy of economists’ forecasts. The forecast accuracy evaluation encompasses the measurement of economic and property forecasts under the following phases: testing for the forecast accuracy; analysing outliers of forecast errors; and testing of causal relationships. Forecast accuracy measurement incorporates scale independent metrics that include Theil’s U values (U1 and U2) and mean absolute scaled error. Inter-quartile range rule is used for the outlier analysis. To find the causal relationships among variables, the time series regression methodology is utilised, including multiple regression analysis and Granger causality developed under the vector auto regression (VAR). Findings The credibility of economic and property forecasts was questionable around the period of the Global Financial Crisis (GFC); a significant man-made Black Swan event. The forecast accuracy measurement highlighted rental movement and net absorption forecast errors as the critical inaccurate predictions. These key property variables are explained by historic information and independent economic variables. However, these do not explain the changes when error time series of the variables were concerned. According to VAR estimates, all property variables have a significant causality derived from the lagged values of Australian S&P/ASX 200 (ASX) forecast errors. Therefore, lagged ASX forecast errors could be used as a warning signal to adjust property forecasts. Research limitations/implications Secondary data were obtained from the premier Australian property markets: Canberra, Sydney, Brisbane, Adelaide, Melbourne and Perth. A limited ten-year timeframe (2001-2011) was used in the ex-post analysis for the comparison of economic and property variables. Forecasts ceased from 2011, due to the discontinuity of the Australian Financial Review quarterly survey of economists; the main source of economic forecast data. Practical implications The research strongly recommended naïve forecasts for the property variables, as an input determinant in each office market forecast equation. Further, lagged forecast errors in the ASX could be used as a warning signal for the successive property forecast errors. Hence, data adjustments can be made to ensure the accuracy of the Australian office market forecasts. Originality/value The paper highlights the critical inaccuracy of the Australian office market forecasts around the GFC. In an environment of increasing incidence of unknown events, these types of risk events should not be dismissed as statistical outliers in real estate modelling. As a proactive strategy to improve office market forecasts, lagged ASX forecast errors could be used as a warning signal. This causality was mirrored in rental movements and total vacancy forecast errors. The close interdependency between rents and vacancy rates in the forecasting process and the volatility in rental cash flows reflects on direct property investment and subsequently on the ASX, is therefore justified.

Warm-sector heavy rainfall in southern China refers to the heavy rainfall that occurs within the warm sector hundreds of kilometers south of a front or without a front during April–June, characterized by poor predictability and a close relationship with low-level jets (LLJs). Based on 45 warm-sector heavy rainfall episodes in 2013 and 2014 in southern China, this study examines their general characteristics and evaluates the performance of convection-permitting WRF Model simulations from an LLJ perspective. The results show that 64% of the warm-sector heavy rainfall episodes are associated with an LLJ (LLJ type) and 36% are not (no-LLJ type). The LLJ type is distinct from the no-LLJ type, with large rainfall accumulation along the coastal area. It is more common for LLJs to occur at both 800 and 925 hPa in the LLJ type, where there is a wide 800-hPa LLJ west of Guangdong Province and two 925-hPa LLJs over Beibu Gulf and the South China Sea (SCS). The coastal convergence associated with the terminus of the LLJ on 925 hPa is conducive to the coastal rainfall. WRF generally presents lower QPF skill in the LLJ type than in the no-LLJ type, due to the severe underestimation of coastal rainfall. The QPF skill of the LLJ type is significantly correlated with the forecast accuracy of LLJs, especially at 925 hPa. The north bias of the simulated LLJ on 925 hPa over the SCS and the associated overestimation of wind speed below ~900 hPa over the inland region weaken the coastal convergence and eventually lead to the underestimation in coastal precipitation.

Solar energy generated from photovoltaic panel is an important energy source that brings many benefits to people and the environment. This is a growing trend globally and plays an increasingly important role in the future of the energy industry. However, it intermittent nature and potential for distributed system use require accurate forecasting to balance supply and demand, optimize energy storage, and manage grid stability. In this study, 5 machine learning models were used including: Gradient Boosting Regressor (GB), XGB Regressor (XGBoost), K-neighbors Regressor (KNN), LGBM Regressor (LightGBM), and CatBoost Regressor (CatBoost). Leveraging a dataset of 21045 samples, factors like Humidity, Ambient temperature, Wind speed, Visibility, Cloud ceiling and Pressure serve as inputs for constructing these machine learning models in forecasting solar energy. Model accuracy is meticulously assessed and juxtaposed using metrics such as coefficient of determination (R 2 ), Root Mean Square Error (RMSE), and Mean Absolute Error (MAE). The results show that the CatBoost model emerges as the frontrunner in predicting solar energy, with training values of R 2 value of 0.608, RMSE of 4.478 W and MAE of 3.367 W and the testing value is R 2 of 0.46, RMSE of 4.748 W and MAE of 3.583 W. SHAP analysis reveal that ambient temperature and humidity have the greatest influences on the value solar energy generated from photovoltaic panel.

This paper studies information sharing in a distribution channel where the manufacturer possesses better demand-forecast information than the downstream retailer. We examine three information-sharing formats: no information sharing (i.e., the manufacturer ex ante commits to not sharing its forecast), voluntary information sharing (i.e., the manufacturer makes the sharing decision ex post after receiving the forecast), and mandatory information sharing (i.e., the manufacturer is mandated to share its forecast). We characterize the equilibrium outcomes under the three sharing formats and investigate the firms’ preferences regarding these formats. It is shown that when the retailer is risk-neutral, both firms are indifferent between voluntary and mandatory sharing. Among the three formats, ex ante, the retailer prefers the no-sharing format whereas the manufacturer prefers the mandatory-sharing format. In addition, we find that a more accurate forecast benefits both firms under voluntary- and mandatory-sharing formats, but may hurt both firms under the no-sharing format. Finally, we show that risk aversion plays a critical role in the firms’ sharing decisions and the impact of forecast accuracy. Specifically, when the retailer is risk-averse, the manufacturer may prefer the no-sharing format over the voluntary-sharing format, and improving forecast accuracy may hurt both firms even under voluntary sharing.

We examine whether exogenous and extremely negative events, such as terrorist attacks and mass shootings, influence the sentiment and forecasts of sell-side equity analysts. We find that analysts who are local to these attacks issue forecasts that are relatively more pessimistic than the consensus forecast. This effect is stronger when the analyst is closer to the event and located in a low-crime region. Impacted analysts are also relatively more pessimistic around the one- and two-year anniversaries of the attacks. Collectively, these findings indicate that exposure to extreme negative events affects the behavior of information intermediaries and the information dissemination process in financial markets. This paper was accepted by David Simchi-Levi, finance.

The data‐driven approaches for medium‐range weather forecasting are recently shown to be extraordinarily promising for ensemble forecasting due to their fast inference speed compared to the traditional numerical weather prediction models. However, their forecast accuracy can hardly match the state‐of‐the‐art operational ECMWF Integrated Forecasting System (IFS) model. Previous data‐driven approaches perform ensemble forecasting using some simple perturbation methods, like the initial condition perturbation and the Monte Carlo dropout. However, their ensemble performance is often limited arguably by the sub‐optimal ways of applying perturbation. We propose a Swin Transformer‐based Variational Recurrent Neural Network (SwinVRNN), which is a stochastic weather forecasting model combining a SwinRNN predictor with a perturbation module. SwinRNN is designed as a Swin Transformer‐based recurrent neural network, which predicts the future states deterministically. Furthermore, to model the stochasticity in the prediction, we design a perturbation module following the Variational Auto‐Encoder paradigm to learn the multivariate Gaussian distributions of a time‐variant stochastic latent variable from the data. Ensemble forecasting can be easily performed by perturbing the model features leveraging the noise sampled from the learned distribution. We also compare four categories of perturbation methods for ensemble forecasting, that is, fixed distribution perturbation, learned distribution perturbation, MC dropout, and multi model ensemble. Comparisons on the WeatherBench data set show that the learned distribution perturbation method using our SwinVRNN model achieves remarkably improved forecasting accuracy and reasonable ensemble spread due to the joint optimization of the two targets. More notably, SwinVRNN surpasses operational IFS on the surface variables of the 2‐m temperature and the 6‐hourly total precipitation at all lead times up to 5 days (Code is available at https://github.com/tpys/wwprediction). Plain Language Summary Ensemble forecasting plays a crucial role in numerical weather prediction (NWP), since a single deterministic model is hard to forecast the chaotic atmosphere conditions. Recent works have begun to explore the data‐driven based ensemble methods due to their rapid prediction speed over traditional NWP. We develop an efficient and effective deep learning model capable of generating large ensemble forecasts with high prediction accuracy and low prediction time cost. The predicted ensemble members have much greater and more reasonable ensemble spread, and better coverage of the ground truth, compared to the prior data‐driven methods. Moreover, our model surpasses the state‐of‐the‐art operational NWP model on the surface atmospheric variables of the 2‐m temperature and the 6‐hourly total precipitation, offering an impressive probability weather prediction baseline. Key Points A transformer‐based variational model called SwinVRNN is developed for medium‐range weather prediction The proposed SwinVRNN can effectively generate large ensemble forecasts with great prediction accuracy and reasonable ensemble spread The model sets a new state‐of‐the‐art among data‐driven models and surpasses the Integrated Forecast System on key atmospheric variables

In the business sector, it has always been a difficult task to predict the exact daily price of the stock market index; hence, there is a great deal of research being conducted regarding the prediction of the direction of stock price index movement. Many factors such as political events, general economic conditions, and traders' expectations may have an influence on the stock market index. There are numerous research studies that use similar indicators to forecast the direction of the stock market index. In this study, we compare two basic types of input variables to predict the direction of the daily stock market index. The main contribution of this study is the ability to predict the direction of the next day's price of the Japanese stock market index by using an optimized artificial neural network (ANN) model. To improve the prediction accuracy of the trend of the stock market index in the future, we optimize the ANN model using genetic algorithms (GA). We demonstrate and verify the predictability of stock price direction by using the hybrid GA-ANN model and then compare the performance with prior studies. Empirical results show that the Type 2 input variables can generate a higher forecast accuracy and that it is possible to enhance the performance of the optimized ANN model by selecting input variables appropriately.

Firms often make mistakes, from simple manufacturing overruns all the way to catastrophic blunders. However, there is considerable heterogeneity in the nature of corporate responses when faced with evidence that an error has taken place, and, therefore, in the likelihood that such errors will reoccur in the future. In this paper, we explore an important but understudied influence on firms' responses to corrective feedback— a CEO's level of overconfidence. Using multiple distinct measures of overconfidence and the empirical context of voluntary corporate earnings forecasts, we find strong, robust evidence that firms led by overconfident CEOs are less responsive to corrective feedback in improving management forecast accuracy. We further show that this relationship is moderated by prior forecast error valence, time horizon, and managerial discretion.