|
on Forecasting |
| By: | Sonan Memon (Pakistan Institute of Development Economics) |
| Abstract: | I begin by motivating the utility of high-frequency inflation estimation and reviewing recent work done at the State Bank of Pakistan for inflation forecasting and now-casting GDP using machine learning (ML) tools. I also present stylised facts about the structure of historical and especially recent inflation trends in Pakistan. |
| Keywords: | Forecast Accuracy, Forecasts of Inflation in Pakistan, High Frequency, Hyperinflation, Inflation Estimation and Forecasting, Machine Learning, Synthetic Data, VAR Models, Web Scrapping and Scanner Data, |
| JEL: | C53 E30 E31 E32 E37 E47 E52 E58 |
| Date: | 2022 |
| URL: | http://d.repec.org/n?u=RePEc:pid:wpaper:2022:12&r=for |
| By: | Mostafa R. Sarkandiz |
| Abstract: | Throughout the past year, Turkey's central bank policy to decrease the nominal interest rate has caused episodes of severe fluctuations in Turkish lira exchange rates. According to these conditions, the daily return of the USD/TRY have attracted the risk-taker investors' attention. Therefore, the uncertainty about the rates has pushed algorithmic traders toward finding the best forecasting model. While there is a growing tendency to employ sophisticated models to forecast financial time series, in most cases, simple models can provide more precise forecasts. To examine that claim, present study has utilized several models to predict daily exchange rates for a short horizon. Interestingly, the simple exponential smoothing model outperformed all other alternatives. Besides, in contrast to the initial inferences, the time series neither had structural break nor exhibited signs of the ARCH and leverage effects. Despite that behavior, there was undeniable evidence of a long-memory trend. That means the series tends to keep a movement, at least for a short period. Finally, the study concluded the simple models provide better forecasts for exchange rates than the complicated approaches. |
| Date: | 2023–02 |
| URL: | http://d.repec.org/n?u=RePEc:arx:papers:2302.08897&r=for |
| By: | Francis X. Diebold (University of Pennsylvania); Maximilian Gobel (University of Lisbon); Philippe Goulet Coulombe (University of Quebec in Montreal) |
| Abstract: | We use "glide charts" (plots of sequences of root mean squared forecast errors as the target date is approached) to evaluate and compare fixed-target forecasts of Arctic sea ice. We first use them to evaluate the simple feature-engineered linear regression (FELR) forecasts of Diebold and Göbel (2022), and to compare FELR forecasts to naive pure-trend benchmark forecasts. Then we introduce a much more sophisticated feature-engineered machine learning (FEML) model, and we use glide charts to evaluate FEML forecasts and compare them to a FELR benchmark. Our substantive results include the frequent appearance of predictability thresholds, which differ across months, meaning that accuracy initially fails to improve as the target date is approached but then increases progressively once a threshold lead time is crossed. Also, we find that FEML can improve appreciably over FELR when forecasting "turning point" months in the annual cycle at horizons of one to three months ahead. |
| Keywords: | Seasonal climate forecasting, forecast evaluation and comparison, prediction |
| JEL: | Q54 C22 C52 C53 |
| Date: | 2022–07 |
| URL: | http://d.repec.org/n?u=RePEc:bbh:wpaper:22-04&r=for |
| By: | Maudud Hassan Uzzal (University of Warsaw, Faculty of Economic Sciences, Quantitative Finance Research Group); Robert Ślepaczuk (University of Warsaw, Faculty of Economic Sciences, Department of Quantitative Finance, Quantitative Finance Research Group) |
| Abstract: | Based on one step ahead forecasts, this study compares the forecasting abilities of the traditional technique (ARIMA) with recurrent neural network (LSTM). In order to check the possible use of these forecasts in different asset management methods, these forecasts are afterwards included into trading signals of investment strategies. As a benchmark method, the Random Walk model producing naive forecasts has been utilized. This research examines daily data from the S&P 500 index for 20 years, from 2000 to 2020, and it includes information on some significant market turbulence. The methods were tested in terms of robustness to changes in parameters and hyperparameters and evaluated based on various error metrics (MAE, MAPE, RMSE MSE). The results show that ARIMA outperforms LSTM in terms of one step ahead forecasts. Finally, LSTM model with a variety of hyperparameters - including a number of epochs, a loss function, an optimizer, activation functions, a number of units, a batch size, and a learning rate - was tested in order to check its robustness. |
| Keywords: | deep learning, recurrent neural networks, ARIMA, algorithmic investment strategies, trading systems, LSTM, walk-forward process, optimization |
| JEL: | C4 C14 C45 C53 C58 G13 |
| Date: | 2023 |
| URL: | http://d.repec.org/n?u=RePEc:war:wpaper:2023-05&r=for |