|
on Forecasting |
| By: | Maheronnaghsh, Mohammad Javad; Gheidi, Mohammad Mahdi; Younesi, Abolfazl; Fazli, MohammadAmin |
| Abstract: | In the dynamic world of financial markets, accurate price predictions are essential for informed decision-making. This research proposal outlines a comprehensive study aimed at forecasting stock and currency prices using state-of-the-art Machine Learning (ML) techniques. By delving into the intricacies of models such as Transformers, LSTM, Simple RNN, NHits, and NBeats, we seek to contribute to the realm of financial forecasting, offering valuable insights for investors, financial analysts, and researchers. This article provides an in-depth overview of our methodology, data collection process, model implementations, evaluation metrics, and potential applications of our research findings. The research indicates that NBeats and NHits models exhibit superior performance in financial forecasting tasks, especially with limited data, while Transformers require more data to reach full potential. Our findings offer insights into the strengths of different ML techniques for financial prediction, highlighting specialized models like NBeats and NHits as top performers - thus informing model selection for real-world applications. To enhance readability, all acronyms used in the paper are defined below: ML: Machine Learning LSTM: Long Short-Term Memory RNN: Recurrent Neural Network NHits: Neural Hierarchical Interpolation for Time Series Forecasting NBeats: Neural Basis Expansion Analysis for Time Series ARIMA: Autoregressive Integrated Moving Average GARCH: Generalized Autoregressive Conditional Heteroskedasticity SVMs: Support Vector Machines CNNs: Convolutional Neural Networks MSE: Mean Squared Error MAE: Mean Absolute Error RMSE: Recurrent Mean Squared Error API: Application Programming Interface F1-score: F1 Score GRU: Gated Recurrent Unit yfinance: Yahoo Finance (a Python library for fetching financial data) |
| Date: | 2023–09–30 |
| URL: | https://d.repec.org/n?u=RePEc:osf:osfxxx:dzp26_v1 |
| By: | Xiaohu Wang (School of Economics, Fudan University, Shanghai, China); Weilin Xiao (School of Management, Zhejiang University, Hangzhou, 310058, China); Jun Yu (Faculty of Business Administration, University of Macau, Macao, China); Chen Zhang (Faculty of Business Administration, University of Macau, Macao, China) |
| Abstract: | This paper first derives two analytic formulae for the autocovariance of the discretely sampled fractional Ornstein-Uhlenbeck (fOU) process. Utilizing the analytic formulae, two main applications are demonstrated: (i) investigation of the accuracy of the likelihood approximation by the Whittle method; (ii) the optimal forecasts with fOU based on discretely sampled data. The finite sample performance of the Whittle method and the derived analytic formula motivate us to introduce a feasible exact maximum likelihood (ML) method to estimate the fOU process. The long-span asymptotic theory of the ML estimator is established, where the convergence rate is a smooth function of the Hurst parameter (i.e., H) and the limiting distribution is always Gaussian, facilitating statistical inference. The asymptotic theory is different from that of some existing estimators studied in the literature, which are discontinuous at H = 3/4 and involve non-standard limiting distributions. The simulation results indicate that the ML method provides more accurate parameter estimates than all the existing methods, and the proposed optimal forecast formula offers a more precise forecast than the existing formula. The fOU process is applied to fit daily realized volatility (RV) and daily trading volume series. When forecasting RVs, it is found that the forecasts generated using the optimal forecast formula together with the ML estimates outperform those generated from all possible combinations of alternative estimation methods and alternative forecast formula. |
| Keywords: | Fractional Ornstein-Uhlenbeck process; Hurst parameter; Out-of-sample forecast; Maximum likelihood; Whittle likelihood; Composite likelihood |
| JEL: | C15 C22 C32 |
| Date: | 2025–03 |
| URL: | https://d.repec.org/n?u=RePEc:boa:wpaper:202527 |
| By: | Douglas Kiarelly Godoy de Araujo; Nikola Bokan; Fabio Alberto Comazzi; Michele Lenza |
| Abstract: | Word embeddings are vectors of real numbers associated with words, designed to capture semantic and syntactic similarity between the words in a corpus of text. We estimate the word embeddings of the European Central Bank's introductory statements at monetary policy press conferences by using a simple natural language processing model (Word2Vec), only based on the information and model parameters available as of each press conference. We show that a measure based on such embeddings contributes to improve core inflation forecasts multiple quarters ahead. Other common textual analysis techniques, such as dictionary-based metrics or sentiment metrics do not obtain the same results. The information contained in the embeddings remains valuable for out-of-sample forecasting even after controlling for the central bank inflation forecasts, which are an important input for the introductory statements. |
| Keywords: | embeddings, inflation, forecasting, central bank texts |
| JEL: | E31 E37 E58 |
| Date: | 2025–03 |
| URL: | https://d.repec.org/n?u=RePEc:bis:biswps:1253 |
| By: | Francis X. Diebold (University of Pennsylvania & NBER); Aaron Mora (University of South Carolina); Minchul Shin (Federal Reserve Bank of Philadelphia) |
| Abstract: | We study the properties of macroeconomic survey forecast response averages as the number of survey respondents grows. Such averages are “portfolios” of forecasts. We characterize the speed and pattern of the gains from diversification and their eventual decrease with portfolio size (the number of survey respondents) in both (1) the key real-world data-based environment of the U.S. Survey of Professional Forecasters (SPF), and (2) the theoretical model-based environment of equicorrelated forecast errors. We proceed by proposing and comparing various direct and model-based “crowd size signature plots”, which summarize the forecasting performance of k-average forecasts as a function of k, where k is the number of forecasts in the average. We then estimate the equicorrelation model for growth and inflation forecast errors by choosing model parameters to minimize the divergence between direct and model-based signature plots. The results indicate near-perfect equicorrelation model fit for both growth and inflation, which we explicate by showing analytically that, under conditions, the direct and fitted equicorrelation model-based signature plots are identical at a particular model parameter configuration. We find that the gains from diversification are greater for inflation forecasts than for growth forecasts, but that both gains nevertheless decrease quite quickly, so that fewer SPF respondents than currently used may be adequate. |
| Keywords: | Macroeconomic surveys of professional forecasters, forecast combination, model averaging, equicorrelation |
| JEL: | C5 C8 E3 E6 |
| Date: | 2025–03–13 |
| URL: | https://d.repec.org/n?u=RePEc:pen:papers:25-008 |
| By: | Vasilios Plakandaras (Department of Economics, Democritus University of Thrace, Komotini, Greece); Rangan Gupta (Department of Economics, University of Pretoria, Private Bag X20, Hatfield 0028, South Africa); Qiang Ji (Institutes of Science and Development, Chinese Academy of Sciences, Beijing, China; School of Public Policy and Management, University of Chinese Academy of Sciences, Beijing, 100049, China) |
| Abstract: | The study investigates systemic financial risk in global markets, attributing it to geopolitical instability, climate risks, and economic uncertainties. Utilizing a state-of-the-art machine learning heterogeneous panel regression framework capable of capturing cross-sectional dependencies and nonlinear patterns, we examine financial stress across multiple economies, including China, the U.S., the U.K., and ten EU nations. Through extensive out-of-sample rolling window analysis, we show that while geopolitical uncertainty enhances short-term predictions, long-term risk forecasting is better achieved using financial and economic data. The study underscores the limitations of conventional regression models in capturing financial risk dynamics and suggests that machine learning-based panel regressions provide a more nuanced and accurate forecasting tool. The findings bear significant policy implications, highlighting the necessity for regulatory bodies to reassess risk frameworks and the role of climate-related disclosures in financial markets. |
| Keywords: | Systemic financial risk, machine learning, forecasting, climate risk, geopolitical risk |
| JEL: | C45 C58 G17 |
| Date: | 2025–03 |
| URL: | https://d.repec.org/n?u=RePEc:pre:wpaper:202511 |
| By: | Luca Benati |
| Abstract: | I use Bayesian VARs with stochastic volatility to forecast global temperatures and sea level and ice cover in the Northern hemisphere until 2100, by exploiting (i) their long-run equilibrium relationship with climate change drivers (CCDs) and (ii) the relationship between world GDP and anthropogenic CCDs. Assuming that trend GDP growth will remain unchanged after 2024, and the world economy will fully decarbonize by 2050, global temperatures and sea level are projected to increase by 2.3 Celsius degrees and 38 centimeters respectively compared to pre-industrial times. Further, uncertainty is substantial, pointing to significant upward risks. Because of this, bringing climate change under control will require massive programme of carbon removal from the atmosphere, in order to bring anthropogenic CCDs back to the levels of the end of the XX century. |
| Keywords: | Climate Change; Bayesian VARs; stochastic volatility; cointegration; forecasting; conditional forecasts |
| JEL: | E2 E3 |
| Date: | 2025–03 |
| URL: | https://d.repec.org/n?u=RePEc:ube:dpvwib:dp2503 |
| By: | Corinna Ghirelli (BANCO DE ESPAÑA); Javier J. Pérez (BANCO DE ESPAÑA); Daniel Santabárbara (BANCO DE ESPAÑA) |
| Abstract: | This paper investigates the relationship between inflation and GDP growth forecast errors and the expected monetary policy stance in the euro area during the monetary policy cycle of 2022-2024, when inflation was well above the ECB’s target. Under rational expectations, forecasts of monetary contractions should be unrelated to subsequent inflation and growth forecast errors. On the contrary, we find that expected monetary policy tightening has been associated with higher than projected GDP growth, suggesting a lower monetary policy effect than that factored in by (ECB/Eurosystem and IMF) forecasters. In other words, forecasters overestimated the monetary multiplier. At the same time, monetary policy tightening has been associated with lower than expected inflation, suggesting an underestimation of the monetary multiplier on inflation. Putting these two stylized facts together implies that forecasters overestimated the sacrifice ratio during the last monetary policy tightening cycle. Our findings suggest that forecasters may have inaccurately perceived the recent inflationary crisis in the euro area as predominantly supply-driven, underestimating its demand-driven component. This led to the belief that monetary policy in the euro area would be exceedingly costly in terms of output. |
| Keywords: | forecast errors, monetary policy multipliers, sacrifice ratio |
| JEL: | C53 E27 E62 E52 E58 |
| Date: | 2025–03 |
| URL: | https://d.repec.org/n?u=RePEc:bde:wpaper:2516 |
| By: | Jonas D. M. Fisher; Leonardo Melosi; Sebastian Rast |
| Abstract: | Professional forecasters’ long-run inflation expectations overreact to news and exhibit persistent, predictable biases in forecast errors. A model incorporating overconfidence in private information and a persistent expectations bias—which generates persistent forecast errors across most forecasters—accounts for these two features of the data, offering a valuable tool for studying long-run inflation expectations. Our analysis highlights substantial, time- varying heterogeneity in forecasters’ responses to public information, with sensitivity declining across all forecasters when monetary policy is constrained by the effective lower bound. The model provides a framework to evaluate whether policymakers’ communicated inflation paths are consistent with anchored long-run expectations. |
| Keywords: | Panel survey data; long-run inflation expectations; rationality; expectation bias; overconfidence; overreaction; central bank communications; anchoring |
| JEL: | E31 D83 E52 E37 |
| Date: | 2025–03 |
| URL: | https://d.repec.org/n?u=RePEc:dnb:dnbwpp:829 |