Journal of Forecasting最新文献

{"title":"The Impact of News Sentiment on the Bitcoin Price via Machine Learning and Deep Learning-Based NLP Models","authors":"Yunus Emre Gür,&nbsp;Emre Ünal","doi":"10.1002/for.70068","DOIUrl":"https://doi.org/10.1002/for.70068","url":null,"abstract":"<div>\u0000 \u0000 <p>This paper employs deep learning and machine learning-based NLP models to investigate the impact of the news sentiment on the Bitcoin price. The lagged Bitcoin variables, news indicators, macroeconomic, and financial factors were taken into account to explain the importance of news sentiment on the Bitcoin price. Moreover, FinBERT-based sentiment scores and semantic features extracted from over 650,000 financial news headlines were integrated with financial and macroeconomic variables. The importance scores of the investigation showed that Bitcoin was largely explained by its lagged price movements, which suggests the speculative nature of the cryptocurrency. However, the investigation also revealed that Bitcoin was significantly influenced by the news sentiment score. In other words, the paper indicates that the movements in the Bitcoin price can be predominantly explained by the news sentiment. Advanced hybrid models (all ML and DL models with the addition of variables obtained with the FinBERT model) were optimized using Optuna and RandomizedSearchCV. The FinBERT-LSTM model achieved the best prediction accuracy. Nevertheless, the main findings indicated that the response of the Bitcoin price to negative news was much stronger than to positive and neutral news. This finding suggests that the asymmetric relationship between the Bitcoin price and news sentiment was evident. GARCH-based volatility and what-if scenario analyses further demonstrated that negative sentiment leads to sharper fluctuations in the Bitcoin price. The paper provides important implications for policymakers, portfolio managers, investors, and academics.</p>\u0000 </div>","PeriodicalId":47835,"journal":{"name":"Journal of Forecasting","volume":"45 3","pages":"895-923"},"PeriodicalIF":2.7,"publicationDate":"2026-03-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147567541","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"经济学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A Universal Kriging Predictor for Probability Density Function Based on Gaussian Mixture Model","authors":"Lei Qin,&nbsp;Yinzhi Wang,&nbsp;Yingqiu Zhu,&nbsp;Ben-Chang Shia","doi":"10.1002/for.70086","DOIUrl":"https://doi.org/10.1002/for.70086","url":null,"abstract":"<div>\u0000 \u0000 <p>With an increase in the demand of infinite-dimensional data analysis, the research community has focused on the analysis of probability density distributions (PDFs). The modeling of PDF is a critical issue in many applications, for example, inflation rate and income distribution. In spatial analysis, making statistical inferences at unobserved locations is a critical task. Many interpolation methods, such as the Kriging method, are developed to appropriately address this problem. However, if we aim to infer PDFs at unobserved locations, there are very few alternative methods for performing the interpolation of PDFs. To solve this problem, we propose a Kriging interpolation method based on Gaussian mixture models (GMMs) for PDFs. We employ the expectation–maximization (EM) algorithm for estimating the parameters of GMM and utilize a linear solution system to determine the weight coefficients of the Kriging predictor. Furthermore, we conduct a theoretical study of the proposed method and establish the asymptotic normality of parameter estimates. Through extensive simulations, we demonstrate that the proposed method outperforms other existing methods in predicting the PDF at unknown locations. A real-world data analysis based on household income distribution dataset shows that the proposed method is suitable for spatial prediction of PDFs at unknown locations.</p>\u0000 </div>","PeriodicalId":47835,"journal":{"name":"Journal of Forecasting","volume":"45 3","pages":"1188-1202"},"PeriodicalIF":2.7,"publicationDate":"2026-03-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147566200","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"经济学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Robust Real-Time Estimates of the German Output Gap Based on a Multivariate Trend-Cycle Decomposition","authors":"Tino Berger,&nbsp;Christian Ochsner","doi":"10.1002/for.70079","DOIUrl":"https://doi.org/10.1002/for.70079","url":null,"abstract":"<div>\u0000 \u0000 <p>The German economy is an important economic driver in the Euro area in terms of gross domestic product, labor force, and international integration. We provide a state of the art estimate of the German output gap between 1995 and 2022 and present a nowcasting scheme that accurately predicts the German output gap up to 3 months prior to a gross domestic product data release. To this end, we elicit a mixed-frequency Bayesian vector-autoregressive model (MF-BVAR) using monthly information to form an expectations about the current-quarter output gap. The mean absolute error of the MF-BVAR nowcast compared to the final estimate is very small (0.28 percentage points) after only 1 month of observed data. Moreover, we show that business and consumer expectations, international trade, and labor market aggregates consistently explain large shares of variation in the German output gap. Finally, the MF-BVAR procedure is very reliable, as it implies an output gap that is hardly revised ex post. This is particularly important for policymakers.</p>\u0000 </div>","PeriodicalId":47835,"journal":{"name":"Journal of Forecasting","volume":"45 3","pages":"1129-1144"},"PeriodicalIF":2.7,"publicationDate":"2026-03-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147565230","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"经济学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Whether Uncertainty Theory Can Enhance GDP Forecasting From Energy: A New Uncertain MIDAS Model","authors":"Yuxin Shi,&nbsp;Chao Liang,&nbsp;Lu Wang","doi":"10.1002/for.70083","DOIUrl":"https://doi.org/10.1002/for.70083","url":null,"abstract":"<div>\u0000 \u0000 <p>In response to the potential failure of traditional models when faced with issues of nonwhite noise residuals and imprecise data, this study extends the mixed data sampling (MIDAS) model to the field of uncertainty theory to tackle these challenges. Under the framework of uncertainty theory, this research addresses the frequency inconsistency in economic data collection by constructing two types of uncertain MIDAS models, aiming to fill the gap in uncertainty theory's handling of predictive analysis for variables with different frequencies. Furthermore, this study integrates the dual perspectives of energy consumption and energy-related carbon dioxide emissions. By building uncertain MIDAS models with uncertain disturbance terms and traditional MIDAS models, the research systematically establishes and comparatively analyzes univariate and multivariate energy consumption, departmental carbon dioxide emissions, and multivariate models that combine these two perspectives. The study's results not only confirm the nonwhite noise characteristics of residuals and validate the rationality of treating residuals as uncertain disturbance terms but also demonstrate through comparative analysis that the uncertain MIDAS model outperforms the traditional MIDAS model in terms of forecasting effectiveness. Moreover, the multivariate forecasting method that considers both perspectives can more comprehensively describe and predict the US quarterly gross domestic product (GDP), showing its superior predictive capability. Furthermore, by altering the evaluation criterion, substituting GDP with nominal GDP and introducing the control variables for robustness analysis, we have further verified the robustness of the model and its results.</p>\u0000 </div>","PeriodicalId":47835,"journal":{"name":"Journal of Forecasting","volume":"45 3","pages":"1158-1176"},"PeriodicalIF":2.7,"publicationDate":"2026-03-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147566373","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"经济学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"GDP Nowcasting With Artificial Neural Networks: How Much Does Long-Term Memory Matter?","authors":"Kristóf Németh,&nbsp;Dániel Hadházi","doi":"10.1002/for.70061","DOIUrl":"https://doi.org/10.1002/for.70061","url":null,"abstract":"<div>\u0000 \u0000 <p>We apply artificial neural networks (ANNs) to nowcast quarterly GDP growth for the US economy. Using the monthly FRED-MD database, we compare the nowcasting performance of five different ANN architectures: the multilayer perceptron (MLP), the one-dimensional convolutional neural network (1D CNN), the Elman recurrent neural network (RNN), the long short-term memory (LSTM) network, and the gated recurrent unit (GRU). The empirical analysis presents results from two distinctively different evaluation periods. The first (2012:Q1–2019:Q4) is characterized by balanced economic growth, while the second (2012:Q1–2024:Q2) also includes periods of the COVID-19 recession. During the first evaluation period, longer input sequences slightly improve nowcasting performance for some ANNs, but the best accuracy is still achieved with 8-month-long input sequences at the end of the nowcasting window. Results from the second test period depict the role of long-term memory even more clearly. The MLP, the 1D CNN, and the Elman RNN work best with 8-month-long input sequences at each step of the nowcasting window. The relatively weak performance of the gated RNNs also suggests that architectural features enabling long-term memory do not result in more accurate nowcasts for GDP growth. The combined results indicate that the 1D CNN seems to represent a “sweet spot” between the simple time-agnostic MLP and the more complex (gated) RNNs. The network generates nearly as accurate nowcasts as the best competitor for the first test period, while it achieves the overall best accuracy during the second evaluation period. Consequently, as a first in the literature, we propose the application of the 1D CNN for economic nowcasting.</p>\u0000 </div>","PeriodicalId":47835,"journal":{"name":"Journal of Forecasting","volume":"45 3","pages":"924-963"},"PeriodicalIF":2.7,"publicationDate":"2026-03-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147568598","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"经济学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"UK Forecasts of Annual GDP: Their Accuracy and the Information Categories Underlying Their Revisions","authors":"Nigel Meade,&nbsp;Ciaran Driver","doi":"10.1002/for.70071","DOIUrl":"https://doi.org/10.1002/for.70071","url":null,"abstract":"<p>Policy makers are concerned with the accuracy of GDP forecasts and want to understand the reasons for the revision of forecasts. We study these issues by examining forecasts of annual UK GDP growth by a panel of agents, published monthly by HM Treasury. We focus on two main issues: the developing accuracy of the group-mean forecast as horizons shorten and the identification of information categories underlying agents' forecast revisions. The accuracy of the group-mean forecast is poor; there is evidence of information rigidity in forecasts within the target year, and accuracy only improves in May of the target year when contemporary information flows lead to increased accuracy. We find a pessimism bias; the median errors of group-mean forecasts are increasingly positive for horizons shorter than 17 months. We seek to explain revisions to both long- and short-horizon group-mean forecasts and individual agent forecasts. Modeling individual agents' forecast revisions using a moving window, we note a consistent tendency by agents to revise their forecast towards the group-mean. Although their importance varied over time, the main information categories explaining revisions were, over longer horizons, the cost of finance, production, and a business confidence indicator. FX rates and inflation were influential over shorter horizons.</p>","PeriodicalId":47835,"journal":{"name":"Journal of Forecasting","volume":"45 3","pages":"977-996"},"PeriodicalIF":2.7,"publicationDate":"2026-03-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://onlinelibrary.wiley.com/doi/epdf/10.1002/for.70071","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147570068","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"经济学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A Rich-Spatial and Multiscaled Transformer-Based Approach for Long-Term Multivariate Time-Series Forecasting Problem","authors":"Linh Nguyen Thi My,&nbsp;Vu Nguyen,&nbsp;Tham Vo","doi":"10.1002/for.70066","DOIUrl":"https://doi.org/10.1002/for.70066","url":null,"abstract":"<div>\u0000 \u0000 <p>In the era of big data and the Internet of Things (IoT), the rapid growth of online data presents both challenges in management and opportunities for knowledge discovery. Among various data structures, time-series data is widely used across domains such as finance, healthcare, and environmental monitoring. Within this field, multivariate time-series forecasting is crucial for predicting complex, interdependent temporal patterns, making it a key focus for applications like stock market analysis, air quality monitoring, and energy forecasting. Generally speaking, the long-sequential multivariate time-series forecasting task is always considered as challenging as it requires the in-depth capability of sufficiently preserving the joined intradependence and interdependence between the variables within the time-series data in forms of a multichannel learning approach. The long sequence length of multivariate input and predicted data poses a significant challenge for time-series forecasting models, making it difficult to effectively learn temporal and dynamic patterns from historical observations while ensuring accurate long-term predictions. Even with advancements in deep learning (DL), including state-of-the-art transformer-based architecture, the application of multichannel learning for complex multivariate time-series data remains an open research problem. To address this limitation, we propose MCTMF, a novel multichannel transformer-based forecasting technique. Our proposed MCTMF model can assist in extending the series-aware time-series learning framework by incorporating a CNN-based multichannel learning mechanism within the temporal feature encoding process. This enhancement allows MCTMF to effectively capture and model intricate rich-spatial dependencies across multiple variables, significantly improving forecasting accuracy for complex multivariate time-series data. The extensive experiments within real-world multivariate time-series datasets have validated the outperformance of our proposed MCTMF model against the contemporary state-of-the-art transformer-based forecasting models.</p>\u0000 </div>","PeriodicalId":47835,"journal":{"name":"Journal of Forecasting","volume":"45 3","pages":"964-976"},"PeriodicalIF":2.7,"publicationDate":"2026-03-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147570069","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"经济学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A Dynamic Cost-Adjusted AdaCost Model for Credit Prediction of Smallholder Farmers","authors":"XianZhu Shao,&nbsp;YongQiang Du,&nbsp;LuoFei Liang,&nbsp;Xue Xu,&nbsp;Zhiyi Lu","doi":"10.1002/for.70069","DOIUrl":"https://doi.org/10.1002/for.70069","url":null,"abstract":"<div>\u0000 \u0000 <p>Data imbalances constrain credit prediction models. This study proposes a dynamic cost-adjusted AdaCost credit prediction model to improve prediction accuracy. Our model enriches existing research methods on the credit problems of smallholder farmers and opens new avenues for establishing credit prediction models for smallholder farmers. The present study is characterized by two distinctive research innovations: Firstly, it introduces a variable cost-sensitive function that enables adaptively adjusted misclassification costs for each sub-model generated by the AdaCost framework. This advancement effectively addresses the inherent limitation of static cost-sensitive function values in conventional AdaCost models. Secondly, the dynamic nature of the proposed cost-sensitive function induces corresponding variations in both sample weights and model weights within interconnected subsequent sub-models. The mechanism fundamentally resolves the critical oversight in traditional AdaCost methodologies that failed to account for the dynamic interdependencies between cost-sensitive functions and weight adaptation processes. Using data on agricultural loans from a commercial bank in China as empirical data and comparing them with seven baseline models, including AdaBoost, AdaCost, Cost RF, Cost XGBoost, Cost SVM, Cost GBDT, and Cost DT, we found that the proposed dynamic cost-adjusted AdaCost model outperformed the other models. Robustness tests were conducted using two publicly available loan datasets from UCI. They showed that the dynamic cost-adjusted AdaCost model performed better than the AdaBoost and AdaCost models.</p>\u0000 </div>","PeriodicalId":47835,"journal":{"name":"Journal of Forecasting","volume":"45 3","pages":"997-1019"},"PeriodicalIF":2.7,"publicationDate":"2026-03-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147562399","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"经济学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Are the Bank of Korea's Inflation Forecasts Biased Toward the Target?","authors":"Eunkyu Seong,&nbsp;Seojeong Lee","doi":"10.1002/for.70076","DOIUrl":"https://doi.org/10.1002/for.70076","url":null,"abstract":"<div>\u0000 \u0000 <p>The Bank of Korea (BoK) regularly publishes the <i>Economic Outlook</i>, offering forecasts for key macroeconomic variables such as GDP growth, inflation, and unemployment rates. This study examines whether the BoK's inflation forecasts exhibit bias, specifically a tendency to align with its inflation target. We extend the Holden and Peel (1990) test to incorporate state dependency, defining the state of the economy based on whether realized inflation falls below the target at the time of the forecast. Our analysis reveals that the BoK's inflation forecasts are biased under this state-dependent framework. Furthermore, we examine a range of bias correction strategies based on AR(1) and mean error models, including their state-dependent variants. These strategies generally improve forecast accuracy. Among them, the AR(1)-based correction exhibits relatively stable performance, consistently reducing the root mean square error.</p>\u0000 </div>","PeriodicalId":47835,"journal":{"name":"Journal of Forecasting","volume":"45 3","pages":"1092-1109"},"PeriodicalIF":2.7,"publicationDate":"2026-03-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147564389","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"经济学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Exploring the Forecasting of Crude Oil, Gold, and Euro Currency Implied Volatility Indices: Insights From the Decomposed Stock Market Volatility","authors":"Gongyue Jiang,&nbsp;Gaoxiu Qiao,&nbsp;Shiyuan Huang","doi":"10.1002/for.70087","DOIUrl":"https://doi.org/10.1002/for.70087","url":null,"abstract":"<div>\u0000 \u0000 <p>This paper explores whether the information from the stock market can provide positive contents for the implied volatility prediction in the crude oil market, gold market, and foreign exchange market. Specifically, we investigate the predictive effects of realized continuous volatility, realized jump volatility, positive and negative realized semi-variations, and signed jumps from the S&amp;P 500 index on three implied volatility indices, OVX (Crude Oil Volatility Index), GVZ (Gold Volatility Index), and EVZ (Euro Volatility Index). We construct a hybrid method by combining parametric models with machine learning to explore the market spillover effects of stock market information on three markets. The empirical results show that realized measures in the stock market can provide incremental information for the prediction of the implied volatility indices, the positive and negative semi-variations of stock index showing better performance than that of jump volatility. The method of combining FNN with the parametric model shows better performance compared to SVR. The superiority of this hybrid approach is further verified based on the Model Confidence Set test. Furthermore, an economic significance evaluation confirms that the enhanced predictive accuracy translates into significant economic value.</p>\u0000 </div>","PeriodicalId":47835,"journal":{"name":"Journal of Forecasting","volume":"45 3","pages":"1203-1224"},"PeriodicalIF":2.7,"publicationDate":"2026-03-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"147566372","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"经济学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}