Past Lags Research Articles

Statistical Downscaling of SEVIRI Land Surface Temperature to WRF Near-Surface Air Temperature Using a Deep Learning Model

The analysis of the near-surface air temperature is vital for many applications such as urban heat islands and climate change studies. In particular, extreme weather events are typically localized and so should the corresponding adaptation measures. However, climate scientists are often confronted with the difficulty of providing reliable predictions at high spatial resolutions in the order of 1 km. We propose to train a convolutional neural network model to emulate the hourly high-resolution near-surface air temperature field simulated by the Weather Research and Forecasting (WRF) software over a period of 18 months. The model is driven by current and past lags of coarse SEVIRI land surface temperature fields. This nowcasting application implements a downscaling of the spatial resolution of the input by about a factor of four, while establishing a correlation between current and past land surface temperature maps and the current near-surface air temperature field. The U-Net variant that is proposed in this study uses regularization to prevent over-fitting and implements a novel space-time approach, where multiple time steps are fed into the model through 3D convolution layers. Besides LST, the model also uses urban density as additional static input to be able to predict the temperature more accurately in urban areas and to improve the generalizability of the trained model. The performance of our U-Net model is assessed via comparison to an MLR benchmark (ridge regularization). The model is shown to be superior on all performance metrics. It achieves a mean absolute error of 1.36 °C versus 1.49 °C for benchmark (a 9% relative improvement) and a root mean square error of 1.77 °C versus 1.91 °C for benchmark (a 7% relative improvement). The absolute error of the model is less than 2 °C for 77% of the prediction points versus 72% for the benchmark (a 7% relative improvement). The improvement over the benchmark is even more significant during extreme hot periods. We demonstrate the generalizability of the approach by testing the trained model on unseen spatial domains.

Spatial-temporal differences of COVID-19 vaccinations in the U.S.

Although the disparities in COVID-19 outcomes have been proved, they have not been explicitly associated with COVID-19 full vaccinations. This paper examines the spatial and temporal patterns of the county-level COVID-19 case rates, fatality rates, and full vaccination rates in the United States from December 24, 2020 through September 30, 2021. Statistical and geospatial analyses show clear temporal and spatial patterns of the progression of COVID-19 outcomes and vaccinations. In the relationship between two time series, the fatality rates series was positively related to past lags of the case rates series. At the same time, case rates series and fatality rates series were negatively related to past lags of the full vaccination rates series. The lag level varies across urban and rural areas. The results of partial correlation, ordinary least squares (OLS) and Geographically Weighted Regression (GWR) also confirmed that the existing COVID-19 infections and different sets of socioeconomic, healthcare access, health conditions, and environmental characteristics were independently associated with COVID-19 vaccinations over time and space. These results empirically identify the geographic health disparities with COVID-19 vaccinations and outcomes and provide the evidentiary basis for targeting pandemic recovery and public health mitigation actions.

VEST: automatic feature engineering for forecasting

Time series forecasting is a challenging task with applications in a wide range of domains. Auto-regression is one of the most common approaches to address these problems. Accordingly, observations are modelled by multiple regression using their past lags as predictor variables. We investigate the extension of auto-regressive processes using statistics which summarise the recent past dynamics of time series. The result of our research is a novel framework called VEST, designed to perform feature engineering using univariate and numeric time series automatically. The proposed approach works in three main steps. First, recent observations are mapped onto different representations. Second, each representation is summarised by statistical functions. Finally, a filter is applied for feature selection. We discovered that combining the features generated by VEST with auto-regression significantly improves forecasting performance in a database composed by 90 time series with high sampling frequency. However, we also found that there are no improvements when the framework is applied for multi-step forecasting or in time series with low sample size. VEST is publicly available online.

English

Inflation tends to be a relatively persistent process, which means that current and past values should be helpful in forecasting future inflation. Applying this intuition, we construct a basic stochastic model which exploits information embedded in past values of Ghana’s inflation data. Therefore the aim of this study is not to identify the drivers of Ghana’s inflation, but to identify and forecast with the best predicting model for Ghana’s inflation, based on the stochastic mechanisms that governs Ghana’s inflation series. We then use this identified model to forecast one-year-ahead (that is, 2018) inflation using past lags, specifically, monthly inflation, from January 2010 to September 2017. Per our forecast, the Bank of Ghana’s aim of hitting a single digit for the year 2018 will not be realized, even though the year closes with a lower inflation than what it began with. Key words: Inflation, ARIMA model, forecasting, Ghana.

An Empirical Analysis of the Relationship between Economic Growth and Credit Volumes in Hungary

Our research focuses on the impact of the private sector’s credit volume on economic growth. The main purpose of our study is to present the theoretical background between lending activity and economic growth, which is empirically tested for Hungary in 2000-2017. We have used the Vector Autoregression model, where the dependent variable is a linear function of past lags of itself and past lags of the other involved variables. We have identified a significant relationship between private sector credit and general government debt and the pace of economic growth. Based on our empirical research, the base effect, Germany’s economic growth and recession were significant variables.

A Markov-Switching model for building occupant activity estimation

Heating and ventilation strategies in buildings can be improved significantly if information about the current presence and activity level of the occupants is taken into account. Therefore, there is a high demand for inexpensive sensor-based methods to detect the occupancy or occupant activity level. It is shown that the carbon dioxide (CO2) level in a room is dependent on the activity level rather than only on just the number of people. Therefore, this study suggests a new model based on the use of CO2 trajectories to estimate the occupant activity level, trained on measurements both from a school classroom and from a Danish summerhouse. A hidden Markov-switching model was employed to identify the activity level. This modelling approach is a generalization of hidden Markov models, taking autocorrelation in the observations into account. This is done by an additional autoregressive part which models the persistence of the CO2 concentration by relating the current value to past lags. The analysis of one-step prediction residuals shows that this method inherits the dynamics of the CO2 curves much better than an ordinary hidden Markov model, and can therefore be considered a promising candidate for occupant activity estimation. Furthermore, it is shown that the presented model can be used for simulations of activity level and of the accompanying CO2 levels.

Causal nexus between foreign direct investment and economic growth

Purpose– The purpose of this paper is to determine cause and effect relationship between foreign direct investment (FDI) and economic growth (gross domestic product (GDP) taken as proxy) for Brazil, Russia, India, China and South Africa (BRICS nations) individually for the period 1992-2013. Also, the study tries to explore the reasons behind the linkage between FDI and GDP by estimating a linear regression model consisting of both macro-economic and institutional variables.Design/methodology/approach– Johansen cointegration technique followed by vector error correction model (VECM) and standard Granger causality test are employed to investigate the causal linkage between FDI and GDP. To delve into the reasons behind this linkage, an ordinary least square (OLS) technique is also applied to test the linear regression model consisting of net FDI inflows as dependent variable and nine macro- economic and institutional variables. Residual diagnostics is also conducted using Breusch-Godfrey Lagrange Multiplier test for diagnosing the problem of serial correlation, Breusch-Pagan-Godfrey test for examining heteroskedasticity and Jarque Bera test for verifying the normality of residuals.Findings– The Johansen cointegration result establishes a single cointegrating vector (long run relationship) between FDI and GDP for India, China and Brazil. After proving a cointegration, VECM results revealed that there exists unidirectional long run causality running from GDP to FDI in case of Brazil, India and China. Also, it is confirmed that there exists short run causality between FDI and GDP in China, i.e. the past lags of FDI jointly impact the value of GDP. However, for Russia and South Africa, where there is no cointegration in the long run, standard Granger causality test is conducted which reveals that in both the nations, FDI and GDP are independent of one another. The results of OLS technique reveal different country-specific factors causing this linkage between FDI inflows and economic growth.Originality/value– Various researchers in the past have examined this issue of linkage between FDI and GDP in the context of various developing or developed nations. This reveals a gap in the existing literature pertaining to this causal linkage in the context of the BRICS. Thus, this study fills this gap by analyzing not just this causal nexus with the help of VECM and Granger causality techniques but also tries to explore further the reasons for such strong/weak/no link with the help of fitting a regression model which comprises of both macro-economic and institutional country-specific variables influencing this causation.

Cointegration and stock market interdependence: Evidence from South Africa, India and the USA

The purpose of this study is to explore the nature of the association and the possible existence of a shortrun and long-run relationship between the stock-market indices of South Africa, India and the USA. The idea behind this combination is to know how the stock markets of these three prominent countries are related to each other. The study employs monthly data from the stock indices, namely JALSH (South Africa), NIFTY (India) and NASDAQ (USA) composite from April 2004 to March 2014. After testing for the normality of the data distribution and the stationarity of the time series data, this paper discovered a strong correlation between the stock market indices of South Africa, India and the USA. The correlation among the stock markets is high, particularly between South Africa and India. In addition, the paper attempts to discover the presence of any predictive ability among these markets by applying the Granger causality test. The result indicates that the NASDAQ index has no predictive ability as far as the JALSH and NIFTY indices are concerned. However, the JALSH index has a predictive ability on the NIFTY index. After testing the Granger cause relationship, the existence of a long-run and short-run relationship is tested. The long-run relationships among the stock market indices are analysed, following the Johansen and Juselius multivariate cointegration approach. The result suggests the absence of a long-run relationship among the three stock market indices. Short-run relationship is investigated with the Vector Autoregression (VAR) model, and the outcome obtained shows that both the USA and the South African stock markets are predicted only by their own past lags. However, the Indian stock market is seen to be a function of its own past lags and the past lags of the South African stock index.

Forecasting stock returns based on information transmission across global markets using support vector machines

This paper provides evidence that forecasts based on global stock returns transmission yield better returns in day trading, for both developed and emerging stock markets. The study investigates the performance of global stock market price transmission information in forecasting stock prices using support vector regression for six global markets--USA (Dow Jones, S&P500), UK (FTSE-100), India (NSE), Singapore (SGX), Hong Kong (Hang Seng) and China (Shanghai Stock Exchange) over the period 1999---2011. The empirical analysis shows that models with other global market price information outperform forecast models based merely on auto-regressive past lags and technical indicators. Shanghai stock index movement was predicted best by Hang Seng Index opening price (57.69), Hang Seng Index by previous day's S&P500 closing price (54.34), FTSE by previous day's S&P500 closing price (57.94), Straits Times Index by previous day's Dow Jones closing price (54.44), Nifty by HSI opening price (60), S&P500 by STI closing price (55.31) and DJIA by HSI opening price (55.22), and Nifty was found to be the most predictable stock index. Trading using global cues-based forecast model generates greater returns than other models in all the markets. The study provides evidence that stock markets across the globe are integrated and the information on price transmission across markets, including emerging markets, can induce better returns in day trading.

Cointegration and stock market interdependence : evidence from South Africa, India and the USA

The purpose of this study is to explore the nature of the association and the possible existence of a shortrun and long-run relationship between the stock-market indices of South Africa, India and the USA. The idea behind this combination is to know how the stock markets of these three prominent countries are related to each other. The study employs monthly data from the stock indices, namely JALSH (South Africa), NIFTY (India) and NASDAQ (USA) composite from April 2004 to March 2014. After testing for the normality of the data distribution and the stationarity of the time series data, this paper discovered a strong correlation between the stock market indices of South Africa, India and the USA. The correlation among the stock markets is high, particularly between South Africa and India. In addition, the paper attempts to discover the presence of any predictive ability among these markets by applying the Granger causality test. The result indicates that the NASDAQ index has no predictive ability as far as the JALSH and NIFTY indices are concerned. However, the JALSH index has a predictive ability on the NIFTY index. After testing the Granger cause relationship, the existence of a long-run and short-run relationship is tested. The long-run relationships among the stock market indices are analysed, following the Johansen and Juselius multivariate cointegration approach. The result suggests the absence of a long-run relationship among the three stock market indices. Short-run relationship is investigated with the Vector Autoregression (VAR) model, and the outcome obtained shows that both the USA and the South African stock markets are predicted only by their own past lags. However, the Indian stock market is seen to be a function of its own past lags and the past lags of the South African stock index.

Macroeconomic factors and conditional bond volatility: evidence from emerging and developed bond markets

This study investigates the effect of macroeconomic factors on the conditional volatility of developed and emerging bond markets using ARMAGARCH model and examines the effect of macroeconomic factors themselves rather than the effect of announcement of macroeconomic factors. The findings show that the macroeconomic factors exhibit a significant relationship with volatility in all the bond markets, more specifically in the emerging bond markets. It is found that past lags explain bond volatility in India, Brazil, USA, UK and Japan, which reasserts that the assumptions of random walk hypothesis does not hold true and bond markets are predictable in the long run. The predictive power of macroeconomic variables and ARMA terms is high in the Indian and Brazilian bond market compared to the developed bond markets. The study shows that, similar to the equity market, the bond market too incorporates information on economic activity and it is more significant in emerging markets.

Multiple-index approach to multiple autoregressive time series model

Time series which have more than one time dependent variable require building an appropriate model in which the variables not only have relationships with each other, but also depend on previous values in time. Based on developments for a sufficient dimension reduction, we investigate a new class of multiple time series models without parametric assumptions. First, for the dependent and independent time series, we simply use a univariate time series central subspace to estimate the autoregressive lags of the series. Secondly, we extract the successive directions to estimate the time series central subspace for regressors which include past lags of dependent and independent series in a mutual information multiple-index time series. Lastly, we estimate a multiple time series model for the reduced directions. In this article, we propose a unified estimation method of minimal dimension using an Akaike information criterion, for situations in which the dimension for multiple regressors is unknown. We present an analysis using real data from the housing price index showing that our approach is an alternative for multiple time series modeling. In addition, we check the accuracy for the multiple time series central subspace method using three simulated data sets.

Single‐Index Additive Vector Autoregressive Time Series Models

Abstract. We study a new class of nonlinear autoregressive models for vector time series, where the current vector depends on single‐indexes defined on the past lags and the effects of different lags have an additive form. A sufficient condition is provided for stationarity of such models. We also study estimation of the proposed model using P‐splines, hypothesis testing, asymptotics, selection of the order of the autoregression and of the smoothing parameters and nonlinear forecasting. We perform simulation experiments to evaluate our model in various settings. We illustrate our methodology on a climate data set and show that our model provides more accurate yearly forecasts of the El Niño phenomenon, the unusual warming of water in the Pacific Ocean.