High-frequency Time Series Data Research Articles

Modeling and Forecasting High Frequency Currency Exchange Rates. A Comparative Study of ARIMA, ANN, and Hybrid ARIMA-ANN Models

Aims/ Objectives: The study develops comparative results on the modeling and prediction performance of the autoregressive integrated moving average (ARIMA), artificial neural network (ANN) and the hybrid ARIMA-ANN time series models for high frequency data. Methodology: The study made use of ARIMA, ANN, and hybrid ARIMA-ANN models to forecasts the East Africa Community countries' daily currency exchange rates data which were obtained from the Central Bank of Kenya website and covered the period from January 2017 to December 2023. Stationarity of the time series data was established using the ADF test. The Ljung Box test and ACF plots were used to establish and compare the goodness-of-fit of the resultant models while the Mean Absolute Percentage Error (MAPE) and Root Mean Squared Error(RMSE) values were used to compare the prediction performance. Results: The study established that the hybrid ARIMA-ANN methodolgy provided better-fitting models for the currency exchange rates data compared to ARIMA and ANN modeling strategies since all the Lyung Box test statistics had p values greater than 5%. Comparatively, the hybrid methodology registered lower MAPE and RMSE values hence had better prediction accuracy compared to ARIMA and ANN methods. Conclusion: The Hybid methodology improves the modeling and forecasting accuracy over the ARIMA and ANN models for high frequency time series data due to its ability to captures both the linear and nonlinear patterns in the time series data.

Deep Learning Approaches for Predicting Intraday Price Movements: An Evaluation of RNN Variants on High-Frequency Stock Data

This study discusses the comparison of four recurrent neural networks (RNN) models: Simple RNN, Gated Recurrent Unit (GRU), Long Short-Term Memory (LSTM), and Bidirectional RNN (BiRNN), in forecasting minute-level stock price time series data. The performance of these four models is evaluated using the Mean Absolute Percentage Error (MAPE) on a stock dataset from Bank Central Asia (BBCA.JK). The experimental results reveal that the GRU model exhibits the best performance with an average MAPE of 0.0255%, followed by the LSTM model with an average MAPE of 0.0377%. The BiRNN model also demonstrates good performance with an average MAPE of 0.0668%, while the Simple RNN has the highest average MAPE at 0.5118%. This suggests that more complex recurrent architectures like GRU and LSTM have better capabilities in capturing patterns in high-frequency time series data. This study can be expanded by exploring other models such as CNN, conducting tests on diverse datasets, and experimenting with a wider range of hyperparameter variations. Additional variables such as economic indicators, global market data, and social data can also offer a more comprehensive understanding of factors influencing stock prices.

Characteristics and interactions of fluctuation velocities, saltation mass flux, and temperature in the atmospheric surface layer

The essence of wind-blown sand movement is a particle movement system driven by high Reynolds number wall turbulence in the atmospheric surface layer (ASL). Understanding this phenomenon is crucial as the law and mechanism of sand movement have not been effectively revealed at present. In this study, we utilize the high-frequency time series data from the Oceano site, a renowned location for such studies, which includes streamwise wind velocity (u), wall-normal wind velocity (w), temperature (T), and total saltation mass flux (q) to delve deeper into wind-blown sand movement. The results indicate a positive correlation between T and w. Interestingly, this correlation not only escalates with height but also intensifies with the increase of the stratification stability parameters (z/L). Sand influences the intensity ratio of each quadrant to Reynolds stress without altering the time ratio. Both the ejection and sweep processes correlate well with z/L. Similarly, q also exhibits a good correlation with z/L. The buoyancy's impact on the low-frequency fluctuations of q might be through its effect on the low-frequency fluctuations of w, enhancing the sand transport capacity of w. This discovery holds profound implications for the study of two-phase flow in ASL and the precise prediction of surface sediment transport.

Data-driven multi-step energy consumption forecasting with complex seasonality patterns and exogenous variables: Model accuracy assessment in change point neighborhoods

Energy forecasting tools became a significant scope of application for time series modeling due to the specific challenges in energy trading — the forecast of consumption for the whole next trading day based on the limited data availability at the forecast origin. The research article addresses the scope of high-frequency time series data, multiple seasonal patterns, exogenous variables, and nonstationary properties in a multi-step forecast horizon tasks. The contribution of the research is the introduction of a machine- and deep-learning-based data-driven approach for multi-output time series forecasting and mainly an introduction of the new evaluation metric called the Change Point Neighborhood Error (CPNE). The purpose of the metric is to provide a distinctive measure of forecasting accuracy of the proposed models in parts of the time series where a change point or a data drift emerges. The experimental findings indicate a notable improvement in accuracy achieved by machine and deep learning models, resulting in a substantial reduction of the mean absolute percentage error (MAPE) by approximately 45 % compared to the optimal statistical model across both datasets used, and also in terms of change point neighborhood error in comparison to statistical models due to their requirement for stationary data input. Deep learning models may be a viable alternative to machine learning approaches; however, deep learning models require long input sequences for accurate forecasting, whereas machine learning methods require shorter input sequences and can benefit more from feature engineering.

The impact of the Russia-Ukraine War on European Union currencies: a high-frequency analysis

ABSTRACT This article examines how the ongoing Ukraine-Russia war has affected the valuation of five European Union currencies: the Polish Zloty, Hungarian Forint, Czech Koruna, Swedish Krone, and Romanian Leu. Using high-frequency time series data and advanced statistical models, we analyze a 522-week period from 1 January 2013, to 24 February 2023, with 24 February 2022, as a key focal point. The Russia-Ukraine war had varying effects on the currencies of non-Eurozone countries. Poland, Hungary, and Sweden experienced significant impacts, while Czechia and Romania remained relatively stable. Our research also highlights the stability of the Euro as a reserve currency, which attracts investors during times of crisis and protects participating economies from currency fluctuations. Additionally, our analysis underscores the distinct impact of war on these currencies, influenced by their unique political, economic, and financial structures and strengths.

Deep Learning Model for Multivariate High-Frequency Time-Series Data: Financial Market Index Prediction

The stock index is actively used for the realization of profits using derivatives and via the hedging of assets; hence, the prediction of the index is important for market participants. As market uncertainty has increased during the COVID-19 pandemic and with the rapid development of data engineering, a situation has arisen wherein extensive amounts of information must be processed at finer time intervals. Addressing the prevalent issues of difficulty in handling multivariate high-frequency time-series data owing to multicollinearity, resource problems in computing hardware, and the gradient vanishing problem due to the layer stacking in recurrent neural network (RNN) series, a novel algorithm is developed in this study. For financial market index prediction with these highly complex data, the algorithm combines ResNet and a variable-wise attention mechanism. To verify the superior performance of the proposed model, RNN, long short-term memory, and ResNet18 models were designed and compared with and without the attention mechanism. As per the results, the proposed model demonstrated a suitable synergistic effect with the time-series data and excellent classification performance, in addition to overcoming the data structure constraints that the other models exhibit. Having successfully presented multivariate high-frequency time-series data analysis, this study enables effective investment decision making based on the market signals.

NATO’s Expansion and Russia’s Aggressiveness: An Empirical Study from the Perspective of the U.S. Public

Abstract Russia’s aggression on Ukraine has attracted worldwide attention. The relations between Russia’s aggression and NATO’s expansion are being debated with conflicting viewpoints. Based on weekly Google Trends data from 5 March 2017 to 13 February 2022, this study creatively creates a series of time series variables to measure the public agenda in the US. Based on time series modellings, this study finds that the U.S. public believes that there are bi-directional relations between NATO’s expansion and Russia’s aggressiveness. In particular, Russia’s aggressiveness is driven by both NATO’s expansion and Russia’s imperialism. NATO’s expansion is driven by Russia’s foreign policy, Russia’s aggressiveness, and Russia’s imperialism. These conclusions predicted the U.S. government’s policy towards this war. This study contributes to policymakers by providing solid empirical evidence on the U.S. public’s opinion on this important issue and to academia as well by presenting a (still) new quantitative method in international relations. By employing this method, we can leverage readily accessible high-frequency time series data to gain valuable insights into U.S. public opinion, potentially enabling people to predict U.S. foreign policies.

High-Frequency Time Series Data Regression Based on the Functional Bayesian Model Averaging

Regression analysis of high-frequency time series data has been an important problem in the field of statistical learning. In this paper, we propose a novel approach based on multiple basis function expansions and Bayesian model averaging when the predictor variable is a high-frequency time series and the response variable is continuous scalar data. On the one hand, the proposed method avoids the curse of dimensionality and extracts the functional information of the original sequence by transforming the discrete data into continuous functions in a function space consisting of some fixed basis functions. On the other hand, for the choice of basis functions, this paper proposes to adaptively determine the optimal solution using Bayesian model averaging, effectively balancing the variance and bias of the predictive model. The real data analysis shows that the proposed method has smaller mean squared error and absolute error, and is robust compared to other methods. Finally, the proposed method can be further extended to time series data applications such as weather forecasting and stock price prediction.

Identifying Informative Nodes in Attributed Spatial Sensor Networks Using Attention for Symbolic Abstraction in a GNN-based Modeling Approach

Modeling complex data, e.g. time series as well as network-based data, is a prominent area of research. In this paper, we focus on a combination of both, analyzing network-based spatial sensor data which is attributed with high frequency time series information. We apply a symbolic representation and an attention-based local abstraction approach, to enhance interpretability on the respective complex high frequency time series data. For this, we aim at identifying informative measurements captured by the respective nodes of the sensor network. To do so, we demonstrate the efficacy of the Symbolic Fourier Approximation (SFA) and the attention-based symbolic abstraction method to localize relevant node sensor-information, by using a transformer architecture as an encoder for a graph neural network. In our experiments, we compare two seismological datasets to their previous state-of-the-art model, demonstrating the advantages and benefits of our presented approach.

MOVING, THINKING, AND SLEEPING: NOVEL INSIGHTS INTO PHYSICAL AND COGNITIVE HEALTH FROM ACCELEROMETRY DATA

Abstract Physical activity and sleep are well-established predictors of health and longevity with aging. Wrist accelerometers, that produce high-frequency time series data, capture multiple aspects of daily physical activity and sleep 24-hours/day. Historically, the majority of accelerometry-based activity research has employed summary metrics to understand the associations of total daily physical activity and sleep with physical and cognitive health. Although these measures are important for understanding conformity with physical activity and sleep recommendations, they underutilize the potential of these data. Further, the summary metrics may differ by accelerometer type/brand, making it difficult to translate results across device types and studies. This symposium will examine the associations between accelerometry-derived physical activity and various aging-related health outcomes, and compare the measurement properties of two commonly used accelerometers for measuring sleep. Ms. Marino will discuss the association of physical activity volume and fragmentation with the presence of the Apolipoprotein-ε4 genotype in the Baltimore Longitudinal Study of Aging (BLSA), overall and by time of day. Dr. Wanigatunga will present evidence on the association of physical activity patterns with beta amyloid plaques in the BLSA. Dr. Schrack will present the association of physical activity fragmentation and diurnal patterns with peripheral artery disease in the Hispanic Community Health Study/Study of Latinos (HCHS/SOL). Finally, Ms. Liu will compare measurement of sleep variables derived from two commonly used accelerometers. Collectively, these presentations highlight ways to utilize the richness of accelerometry data to illuminate more sensitive associations between movement and health outcomes to advance prevention science and promote health aging.

A high frequency stock price prediction model based on Hermite basis expansion and LSTM neural network

Based on Hermite basis function expansion and ensemble learning, an improved LSTM neural network method is proposed in this paper. The proposed method can be used for high frequency stock price prediction. Considering the characteristics of high frequency time series data such as high dimensionality, large noise and instability, etc. In this paper, the function information extracted from the Hermite basis function expansion is used to predict the residual sequence predicted by the LSTM neural network. Since the components of the function feature vector are unknown to the underlying model structure of the response variable, the proposed method is processed by Bagging framework. It not only captures the structure of the latent model, but also balances the variance and bias of the model. In addition, the number of prediction periods of the LSTM neural network is a hyperparameter, and the model averaging method based on distance covariance weighting is considered in this paper for optimization. The results of actual data analysis show that the proposed method can effectively optimize the prediction accuracy of the LSTM neural network, and has certain robustness. Finally, on the one hand, this optimization framework can be used to improve other time series prediction models. On the other hand, the proposed method can play an important role in forecasting problems such as daily average temperature prediction and real-time monitoring of atmospheric environmental quality.

Temporal Disaggregation of Gross Domestic Product Data with Night-Time Lights Data for South Africa

The absence of high-frequency time series data is a real constraint for many researchers worldwide. Generating high-frequency economic data, especially at the sub-national level, generates significant advantages and opportunities. Research otherwise impossible now becomes possible and indeed desirable. Therefore, this study aims to investigate the usefulness of several temporal disaggregation methods and the reliability of the derived results. Temporal disaggregation is one such method that can be used to generate high-frequency time series data from low-frequency time series data. The methods employed will specifically refer to methods that use related series as the high-frequency time series data related to the South African economy. To this effect, monthly night-time lights sensed by satellite were assessed. The results derived through the study seem to generate robust estimates suggesting that it is indeed possible to use night-time lights as a related indicator and the various temporal disaggregation methods contained within the Ecotrim computer software program and ‘tempdisagg’ R package to disaggregate annual national and provincial gross domestic product (GDP). The derived quarterly and monthly national and provincial GDP time series/data will subsequently assist in addressing the constraints researchers face, especially at a sub-national level. From a policy perspective, policymakers should note that remote sensing technologies offer unique opportunities to conduct sub-national economic policy and research.

Subjective well-being during the 2020-21 global coronavirus pandemic: Evidence from high frequency time series data.

We investigate how subjective well-being varied over the course of the global COVID-19 pandemic, with a special attention to periods of lockdown. We use weekly data from YouGov's Great Britain Mood Tracker Poll, and daily reports from Google Trends, that cover the entire period from six months before until eighteen months after the global spread of COVID-19. Descriptive trends and time-series models suggest that negative mood associated with the imposition of lockdowns returned to baseline within 1-3 weeks of lockdown implementation, whereas pandemic intensity, measured by the rate of fatalities from COVID-19 infection, was persistently associated with depressed affect. The results support the hypothesis that country-specific pandemic severity was the major contributor to increases in negative affect observed during the COVID-19 pandemic, and that lockdowns likely ameliorated rather than exacerbated this effect.

Using dynamic time warping self-organizing maps to characterize diurnal patterns in environmental exposures

Advances in measurement technology are producing increasingly time-resolved environmental exposure data. We aim to gain new insights into exposures and their potential health impacts by moving beyond simple summary statistics (e.g., means, maxima) to characterize more detailed features of high-frequency time series data. This study proposes a novel variant of the Self-Organizing Map (SOM) algorithm called Dynamic Time Warping Self-Organizing Map (DTW-SOM) for unsupervised pattern discovery in time series. This algorithm uses DTW, a similarity measure that optimally aligns interior patterns of sequential data, both as the similarity measure and training guide of the neural network. We applied DTW-SOM to a panel study monitoring indoor and outdoor residential temperature and particulate matter air pollution (PM2.5) for 10 patients with asthma from 7 households near Salt Lake City, UT; the patients were followed for up to 373 days each. Compared to previous SOM algorithms using timestamp alignment on time series data, the DTW-SOM algorithm produced fewer quantization errors and more detailed diurnal patterns. DTW-SOM identified the expected typical diurnal patterns in outdoor temperature which varied by season, as well diurnal patterns in PM2.5 which may be related to daily asthma outcomes. In summary, DTW-SOM is an innovative feature engineering method that can be applied to highly time-resolved environmental exposures assessed by sensors to identify typical diurnal (or hourly or monthly) patterns and provide new insights into the health effects of environmental exposures.

A mutual information based R-vine copula strategy to estimate VaR in high frequency stock market data.

In this paper, we explore mutual information based stock networks to build regular vine copula structure on high frequency log returns of stocks and use it for the estimation of Value at Risk (VaR) of a portfolio of stocks. Our model is a data driven model that learns from a high frequency time series data of log returns of top 50 stocks listed on the National Stock Exchange (NSE) in India for the year 2014. The Ljung-Box test revealed the presence of Autocorrelation as well as Heteroscedasticity in the underlying time series data. Analysing the goodness of fit of a number of variants of the GARCH model on each working day of the year 2014, that is, 229 days in all, it was observed that ARMA(1,1)-EGARCH(1,1) demonstrated the best fit. The joint probability distribution of the portfolio is computed by constructed an R-Vine copula structure on the data with the mutual information guided minimum spanning tree as the key building block. The joint PDF is then fed into the Monte-Carlo simulation procedure to compute the VaR. If we replace the mutual information by the Kendall's Tau in the construction of the R-Vine copula structure, the resulting VaR estimations were found to be inferior suggesting the presence of non-linear relationships among stock returns.

Predicting High-frequency Stock Price Using Machine Learning Technique

This paper addresses the problem of predicting the stock price using the high frequency data based on a machine learning approach. We study two things in this paper (1) comparison of the prediction performance among selected function classes with given look-back parameter in terms of the proposed evaluation measures in the process of finding the best in-sample empirical loss minimizer (2) the comparison of those results by changing the sampled frequency of financial time series data after obtaining an introduced set of high frequency data features extracted from the Trades and Quotes (TAQ) data. For the analysis of TAQ data, feature engineering involves the computation of 56 number of related features including market microstructure, statistical, and technical indicator features. Re-estimation was done to improve the prediction accuracy for data models to obtain the predicted value every moving window. On the other hand, algorithmic models are used without re-estimation for the practical matter in that the time spent to train the model is often larger than the sampled frequency of the data. Moreover, the look-back parameter is introduced to cut off the irrelevant long past historical data. Among the selected function class in the experiment, the results show that the PCA regression performs the best in terms of the mean directional accuracy and simple back testing for both NASDAQ100 index and TAQ data with given sampled frequencies (i.e., 3min, 5min, etc). Compared to previous studies using NASDAQ100, the results demonstrate that re-estimation and properly chosen look-back parameter improve the prediction performance in terms of proposed evaluation measures. When it comes to maximum draw down, which is a measure critical for risk management, DA-RNN rendered the smallest value and, thereby, was the best performing model for TAQ data for all time frequencies. We also provide DM statistics whose null hypothesis is that the accuracies of prediction values of any two given models will not be different. In case of TAQ data for all sampled frequencies, there is evidence that we cannot reject the null hypothesis when comparing between PCA regression and DA-RNN model. Extensive experiments provide insights into properly evaluating the prediction performance of best in-sample empirical loss minimizer using the high frequency time series data.

Measuring Advertising Expenditure Effects on the Nigerian Economy

This study explores relationships between annual advertising expenditure and major macroeconomic variables in Nigeria. Advertising is sometimes viewed as a concern of business units only not worth being researched at macroeconomic level. This nature has been mostly studied on advertising industries in the advanced economies. Due to a lack of high frequency time series data on advertising expenditure in the developing economies, this work has been limited to an exploratory study using the multiple regression and correlation analysis. The study covers the period of 2001 to 2018. Its findings show that advertising has positive relationship with GDP and savings. This study provides further evidence on the cyclical nature of advertising that moves with the state of the economy. During the economic slowdown in the period of 2015 to 2017, Nigeria advertising expenditure continued to fall. In 2013, the period with the highest advertising revenue in the study, the ratio of advertising expenditure as percentage of GDP accounted for 0.061%, which was below 0.2%, a very negligible number indicating more scope for growth in the market.

A framework for automated anomaly detection in high frequency water-quality data from in situ sensors

Monitoring the water quality of rivers is increasingly conducted using automated in situ sensors, enabling timelier identification of unexpected values or trends. However, the data are confounded by anomalies caused by technical issues, for which the volume and velocity of data preclude manual detection. We present a framework for automated anomaly detection in high-frequency water-quality data from in situ sensors, using turbidity, conductivity and river level data collected from rivers flowing into the Great Barrier Reef. After identifying end-user needs and defining anomalies, we ranked anomaly importance and selected suitable detection methods. High priority anomalies included sudden isolated spikes and level shifts, most of which were classified correctly by regression-based methods such as autoregressive integrated moving average models. However, incorporation of multiple water-quality variables as covariates reduced performance due to complex relationships among variables. Classifications of drift and periods of anomalously low or high variability were more often correct when we applied mitigation, which replaces anomalous measurements with forecasts for further forecasting, but this inflated false positive rates. Feature-based methods also performed well on high priority anomalies and were similarly less proficient at detecting lower priority anomalies, resulting in high false negative rates. Unlike regression-based methods, however, all feature-based methods produced low false positive rates and have the benefit of not requiring training or optimization. Rule-based methods successfully detected a subset of lower priority anomalies, specifically impossible values and missing observations. We therefore suggest that a combination of methods will provide optimal performance in terms of correct anomaly detection, whilst minimizing false detection rates. Furthermore, our framework emphasizes the importance of communication between end-users and anomaly detection developers for optimal outcomes with respect to both detection performance and end-user application. To this end, our framework has high transferability to other types of high frequency time-series data and anomaly detection applications.

Stylized Facts of High-frequency Financial Time Series Data

High-frequency financial time series data have an ability to define market microstructure and are helpful in making rational real-time decisions. These data sets carry unique characteristics and properties which are not available in low-frequency data; with that high-frequency data also create more challenges and opportunities for econometric modelling and financial data analysis. So it is essential to know the features and the facts related to the high-frequency time series data. In this article, we provide the characteristics and stylized facts exhibited by the high-frequency financial time series data of the S&P CNX Nifty futures index. Stylized facts are mostly related to the empirical observed behaviours, distributional properties, autocorrelation function and seasonality of the high-frequency data. Also, it illustrates the importance of stationarity in financial time series analysis. The knowledge of such facts and concepts is helpful to establish better empirical models and to produce reliable forecasts.

A New Method to Detect Outliers in High-frequency Time Series

The objective of this research is to develop a fast, simple method for detecting and replacing extreme spikes in high-frequency time series data. The method primarily consists  of a nonparametric procedure that pursues a balance between fidelity to observed data and smoothness. Furthermore, through examination of the absolute difference between original and smoothed values, the technique is also able to detect and, where necessary, replace outliers with less extreme data.
 Unlike other filtering procedures found in the literature, our method does not require a model to be specified for the data. Additionally, the filter makes only a single pass through the time series. Experiments  show that the new method can be validly used as a data preparation tool to ensure that time series modeling is supported by clean data, particularly in a complex context such as one with high-frequency data.