Search results

Better forecasting always leads to better management and planning of the operations. The container throughput data are complex and often have multiple seasonality. This makes it difficult to forecast accurately. The purpose of this paper is to forecast container throughput using deep learning methods and benchmark its performance over other traditional time-series methods.

In this study, long short-term memory (LSTM) networks are implemented to forecast container throughput. The container throughput data of the Port of Singapore are used for empirical analysis. The forecasting performance of the LSTM model is compared with seven different time-series forecasting methods, namely, autoregressive integrated moving average (ARIMA), simple exponential smoothing, Holt–Winter’s, error-trend-seasonality, trigonometric regressors (TBATS), neural network (NN) and ARIMA + NN. The relative error matrix is used to analyze the performance of the different models with respect to bias, accuracy and uncertainty.

The results showed that LSTM outperformed all other benchmark methods. From a statistical perspective, the Diebold–Mariano test is also conducted to further substantiate better forecasting performance of LSTM over other counterpart methods.

The proposed study is a contribution to the literature on the container throughput forecasting and adds value to the supply chain theory of forecasting. Second, this study explained the architecture of the deep-learning-based LSTM method and discussed in detail the steps to implement it.

Recently, it has been suggested that macroeconomic forecasts from estimated dynamic stochastic general equilibrium (DSGE) models tend to be more accurate out-of-sample than random walk forecasts or Bayesian vector autoregression (VAR) forecasts. Del Negro and Schorfheide (2013) in particular suggest that the DSGE model forecast should become the benchmark for forecasting horse-races. We compare the real-time forecasting accuracy of the Smets and Wouters (2007) DSGE model with that of several reduced-form time series models. We first demonstrate that none of the forecasting models is efficient. Our second finding is that there is no single best forecasting method. For example, typically simple AR models are most accurate at short horizons and DSGE models are most accurate at long horizons when forecasting output growth, while for inflation forecasts the results are reversed. Moreover, the relative accuracy of all models tends to evolve over time. Third, we show that there is no support to the common practice of using large-scale Bayesian VAR models as the forecast benchmark when evaluating DSGE models. Indeed, low-dimensional unrestricted AR and VAR forecasts may forecast more accurately.

The purpose of this paper is to propose a simple regression‐based method of forecasting daily electricity demand, which may serve as a more accurate benchmark for short‐term forecasts.

In order to make more efficient use of the calendar effects in electricity demand, including weekend, and seasonal effects, while maintaining the parsimony of the forecasting model, the authors match the demand on each day of an entire year with the average of the corresponding days in recent years. This matching‐day approach substantially simplifies the modeling procedure of complex periodicity in electricity demand without loss of information.

With daily data on electric power system load in New England, the authors' method provides quite accurate forecasts. The mean absolute percentage error (MAPE) (2.1 percent) is significantly lower than those of the seasonal ARIMA and exponential smoothing method, and also comparable to the performance of more sophisticated methods in the literature.

The authors' method needs to be modified or augmented by other techniques when the periodicity is not stable due to time trends, economic crises, and other factors.

The management of electric utility providers as well as professional forecasters may use this method as a handy benchmark.

While previous studies focus mainly on accuracy of forecasts, the method presented in the paper is developed with the balance between accuracy and ease of use in mind.

The objective of the paper is to explore the out-of-sample forecasting connections in income growth across the globe.

An autoregressive distributed lag (ARDL) framework is employed and the forecasting performance is analyzed across several horizons using different forecast combination techniques.

Results show that the foreign country's income provides superior forecasts beyond what is provided by the country's own past income movements. Superior forecasting power is particularly held by Belgium, Korea, New Zealand, the UK and the US, while these countries' income is rather difficult to predict by global counterparts. Contrary to conventional wisdom, improved forecasts of income can be obtained even for longer horizons using our approach. Results also show that the forecast combination techniques yield higher forecasting gains relative to individual model forecasts, both in magnitude and the number of countries.

The forecasting paths of income movement across the globe reveal that predictive power greatly differs across countries, regions and forecast horizons. The countries that are difficult to predict in the short run are often seen to be predictable by global income movements in the long run.

Even while it is difficult to predict the income movements at an individual country level, combining information from the income growth of several countries is likely to provide superior forecasting gains. And these gains are higher for long-horizon forecasts as compared to the short-horizon forecast.

In evaluating the forward-looking social implications of economic policy changes, the policymakers should also consider the possible global forecasting connections revealed in the study.

Employing an ARDL model to explore global income forecasting connections across several forecast horizons using different forecast combination techniques.

With the rise in wine consumption, accurate wine price forecasts have significantly impacted restaurant and hotel purchasing decisions and inventory management. This study aims to develop a precise and effective wine price point and interval forecasting model.

The proposed forecast model uses an improved hybrid kernel extreme learning machine with an attention mechanism and a multi-objective swarm intelligent optimization algorithm to produce more accurate price estimates. To the best of the authors’ knowledge, this is the first attempt at applying artificial intelligence techniques to improve wine price prediction. Additionally, an effective method for predicting price intervals was constructed by leveraging the characteristics of the error distribution. This approach facilitates quantifying the uncertainty of wine price fluctuations, thus rendering decision-making by relevant practitioners more reliable and controllable.

The empirical findings indicated that the proposed forecast model provides accurate wine price predictions and reliable uncertainty analysis results. Compared with the benchmark models, the proposed model exhibited superiority in both one-step- and multi-step-ahead forecasts. Meanwhile, the model provides new evidence from artificial intelligence to explain wine prices and understand their driving factors.

This study is a pioneering attempt to evaluate the applicability and effectiveness of advanced artificial intelligence techniques in wine price forecasts. The proposed forecast model not only provides useful options for wine price forecasting but also introduces an innovative addition to existing forecasting research methods and literature.

Online reviews serve as valuable sources of information, reflecting tourists’ attentions, preferences and sentiments. However, although the existing research has demonstrated that incorporating online review data can enhance the performance of tourism demand forecasting models, the reliability of online review data and consumers’ decision-making process have not been given adequate attention. To address the aforementioned problem, the purpose of this study is to forecast tourism demand using online review data derived from the analysis of review helpfulness.

The authors propose a novel “identification-first, forecasting-second” framework. This framework prioritizes the identification of helpful reviews through a comprehensive analysis of review helpfulness, followed by the integration of helpful online review data into the forecasting system. Using the SARIMAX model with helpful online review data sourced from TripAdvisor, this study forecasts tourist arrivals in Hong Kong during the period from August 2012 to June 2019. The SNAÏVE/SARIMA model was used as the benchmark model. Additionally, artificial intelligence models including long short-term memory, back propagation neural network, extreme learning machine and random forest models were used to assess the robustness of the results.

The results demonstrate that online review data are subject to noise and bias, which can adversely affect the accuracy of predictions when used directly. However, by identifying helpful online reviews beforehand and incorporating them into the forecasting process, a notable enhancement in predictive performance can be realized.

First, to the best of the authors’ knowledge, this study is one of the first to focus on the data issue of online reviews on tourism arrivals forecasting. Second, this study pioneers the integration of the consumer decision-making process into the domain of tourism demand forecasting, marking one of the earliest endeavors in this area. Third, this study makes a novel attempt to identify helpful online reviews based on reviews helpfulness analysis.

Empirical publications on the time series modeling and forecasting of electricity prices vary widely regarding the conditions, and the findings make it difficult to generalize results. Against this background, it is surprising that there is a lack of statistics-based literature reviews on the forecasting performance when comparing different models. The purpose of the present study is to fill this gap.

The authors conduct a comprehensive literature analysis from 2000 to 2015, covering 86 empirical studies on the time series modeling and forecasting of electricity spot prices. Various statistics are presented to characterize the empirical literature on electricity spot price modeling, and the forecasting performance of several model types and modifications is analyzed. The key issue of this study is to offer a comparison between different model types and modeling conditions regarding their forecasting performance, which is referred to as a quasi-meta-analysis, i.e. the analysis of analyses to achieve more general findings independent of the circumstances of single studies.

The authors find evidence that generalized autoregressive conditional heteroscedasticity models outperform their autoregressive–moving-average counterparts and that the consideration of explanatory variables improves forecasts.

To the best knowledge of the authors, this paper is the first to apply the methodology of meta-analyses in a literature review of the empirical forecasting literature on electricity spot markets.

Financial market holds superior information that can give insights into the future trajectory of economic growth. Further, identifying sectors that hold the key to future economic growth is important for all economies, but particularly relevant to emerging markets. However, unlike existing studies, the paper provides new insights into the forward-oriented nexus between financial markets and economic growth.

This paper takes a forward-looking approach of using financial market information to predict future economic growth. The authors use ARDL modeling approach to predict economic growth using the information from stock market sectoral returns.

The authors find that sectoral stock returns significantly improve economic growth forecasts. However, the forecasting superiority is not uniform across sectors and horizons. Auto, consumers' spending, materials and realty sectors provide the most forecasting gains. In contrast, banking, capital goods and industrial sectors provide superior forecasts, but only at horizons beyond one year. The authors also find that the forecast superiority of sectors at longer horizons is inversely related to volatility.

Research highlights the need for sector-focused policy actions in driving economic growth. Further, the findings of the paper identify sectors that drive short-, medium- and long-term economic growth.

There is a significant heterogeneity among different sectors and across horizons in predicting economic growth. Results suggest that targeted policy actions in sectors like materials, metals, oil and gas, and realty are key in driving economic growth. Further, policies geared toward the grassroots industries are at least as beneficial as the large-scale industries. Evidence also suggests the need for an active fiscal policy to address infrastructural bottlenecks in primary industries like basic materials and energy. Evidence nevertheless does not undermine the role of monetary policy actions.

Unlike any paper till date, the innovation of the paper is that it takes an ARDL modeling approach to measure stock market sectoral returns' ability to forecast economic growth several months ahead in the future. Though the paper considers the Indian case, the innovation and contribution extents to the entire field of economic studies.

This article surveys recent developments in the evaluation of point and density forecasts in the context of forecasts made by vector autoregressions. Specific emphasis is placed on highlighting those parts of the existing literature that are applicable to direct multistep forecasts and those parts that are applicable to iterated multistep forecasts. This literature includes advancements in the evaluation of forecasts in population (based on true, unknown model coefficients) and the evaluation of forecasts in the finite sample (based on estimated model coefficients). The article then examines in Monte Carlo experiments the finite-sample properties of some tests of equal forecast accuracy, focusing on the comparison of VAR forecasts to AR forecasts. These experiments show the tests to behave as should be expected given the theory. For example, using critical values obtained by bootstrap methods, tests of equal accuracy in population have empirical size about equal to nominal size.

This article explores the importance of accessible and focal information in influencing beliefs and attention in a learning-to-forecast laboratory experiment where subjects are incentivized to form accurate expectations about inflation and the output gap. We consider the effects of salient and accessible forecast error information and learning on subjects’ forecasting accuracy and heuristics, and on aggregate stability. Experimental evidence indicates that, while there is considerable heterogeneity in the heuristics used, subjects’ forecasts can be best described by a constant gain learning model where subjects respond to forecast errors. Salient forecast error information reduces subjects’ overreaction to their errors and leads to greater forecast accuracy, coordination of expectations, and macroeconomic stability. The benefits of this focal information are short-lived and diminish with learning.