High-frequency Time Series Data Research Articles

The Deeds of Speed: An Agent-Based Model of Market Liquidity and Flash Episodes

This paper examines the role of high-frequency traders in flash episodes in electronic financial markets. To do so, we construct an agent-based model of a market for a financial asset in which trading occurs through a central limit order book. The model consists of heterogeneous agents with different trading strategies and frequencies, and is calibrated to high-frequency time series data on the sterling-US dollar exchange rate. Flash episodes occur in the model due to the procyclical behaviour of high-frequency market participants. This is aligned with some empirical evidence as to the drivers of real-world flash crashes. We find that the prevalence of flash episodes increases with the frequency with which high-frequency market participants trade compared to their low-frequency counterparts. This provides tentative theoretical evidence that the recent growth in high-frequency trading across some markets has led to flash episodes. Furthermore, we adapt the model so that large movements in price trigger temporary halts in trading (ie circuit breakers). This is found to reduce the magnitude and frequency of flash episodes.

Intelligent Soft Computing on Forex: Exchange Rates Forecasting with Hybrid Radial Basis Neural Network.

This paper deals with application of quantitative soft computing prediction models into financial area as reliable and accurate prediction models can be very helpful in management decision-making process. The authors suggest a new hybrid neural network which is a combination of the standard RBF neural network, a genetic algorithm, and a moving average. The moving average is supposed to enhance the outputs of the network using the error part of the original neural network. Authors test the suggested model on high-frequency time series data of USD/CAD and examine the ability to forecast exchange rate values for the horizon of one day. To determine the forecasting efficiency, they perform a comparative statistical out-of-sample analysis of the tested model with autoregressive models and the standard neural network. They also incorporate genetic algorithm as an optimizing technique for adapting parameters of ANN which is then compared with standard backpropagation and backpropagation combined with K-means clustering algorithm. Finally, the authors find out that their suggested hybrid neural network is able to produce more accurate forecasts than the standard models and can be helpful in eliminating the risk of making the bad decision in decision-making process.

Characterising phosphorus and nitrate inputs to a rural river using high-frequency concentration–flow relationships

The total reactive phosphorus (TRP) and nitrate concentrations of the River Enborne, southern England, were monitored at hourly interval between January 2010 and December 2011. The relationships between these high-frequency nutrient concentration signals and flow were used to infer changes in nutrient source and dynamics through the annual cycle and each individual storm event, by studying hysteresis patterns. TRP concentrations exhibited strong dilution patterns with increasing flow, and predominantly clockwise hysteresis through storm events. Despite the Enborne catchment being relatively rural for southern England, TRP inputs were dominated by constant, non-rain-related inputs from sewage treatment works (STW) for the majority of the year, producing the highest phosphorus concentrations through the spring–summer growing season. At higher river flows, the majority of the TRP load was derived from within-channel remobilisation of phosphorus from the bed sediment, much of which was also derived from STW inputs. Therefore, future phosphorus mitigation measures should focus on STW improvements. Agricultural diffuse TRP inputs were only evident during storms in the May of each year, probably relating to manure application to land. The nitrate concentration–flow relationship produced a series of dilution curves, indicating major inputs from groundwater and to a lesser extent STW. Significant diffuse agricultural inputs with anticlockwise hysteresis trajectories were observed during the first major storms of the winter period. The simultaneous investigation of high-frequency time series data, concentration–flow relationships and hysteresis behaviour through multiple storms for both phosphorus and nitrate offers a simple and innovative approach for providing new insights into nutrient sources and dynamics.

Hybrid ARIMA/RBF Framework for Prediction BUX Index

In this paper, we construct and implement a new architecture and learning method of customized hybrid RBF neural network for high frequency time series data forecasting. The hybridization is carried out using two running approaches. In the first one, the ARCH (Autoregressive Conditionally Heteroscedastic)-GARCH (Generalized ARCH) methodology is applied. The second modeling approach is based on RBF (Radial Basic Function) neural network using Gaussian activation function with cloud concept. The use of both methods is useful, because there is no knowledge about the relationship between the inputs into the system and its output. Both approaches are merged into one framework to predict the final forecast values. The question arises whether non-linear methods like neural networks can help modeling any non-linearities being inherent within the estimated statistical model. We also test the customized version of the RBF combined with the machine learning method based on SVM learning system. The proposed novel approach is applied to high frequency data of the BUX stock index time series. Our results show that the proposed approach achieves better forecast accuracy on the validation dataset than most available techniques.

Land cover classification of finer resolution remote sensing data integrating temporal features from time series coarser resolution data

Land cover classification of finer resolution remote sensing data is always difficult to acquire high-frequency time series data which contains temporal features for improving classification accuracy. This paper proposed a method of land cover classification with finer resolution remote sensing data integrating temporal features extracted from time series coarser resolution data. The coarser resolution vegetation index data is first fused with finer resolution data to obtain time series finer resolution data. Temporal features are extracted from the fused data and added to improve classification accuracy. The result indicates that temporal features extracted from coarser resolution data have significant effect on improving classification accuracy of finer resolution data, especially for vegetation types. The overall classification accuracy is significantly improved approximately 4% from 90.4% to 94.6% and 89.0% to 93.7% for using Landsat 8 and Landsat 5 data, respectively. The user and producer accuracies for all land cover types have been improved.

In situ measurements of ionospheric plasma turbulence over five frequency decades: Heritage flight of the Plasma Local Anomalous Noise Experiment (PLANE)

Observations of ionospheric plasma density and frequency-dependent broadband plasma turbulence made during the heritage flight of the Plasma Local Anomalous Noise Experiment (PLANE) are presented. Rather than record high frequency time series data, the experiment was designed to record Power Spectral Distributions (PSDs) in five decadal frequency bins with upper limits ranging from 1.0Hz to 10kHz. Additionally, PLANE was designed distinguish turbulence in the ambient plasma from that local to the spacecraft. The instrument consists of two retarding potential analyzers (RPAs) connected together via a feedback loop to force one analyzer into the I–V trace retardation region at all times. Fluctuations in this measurement are believed to be ambient only as the RPA’s voltage would be too high for locally turbulent plasma to surmount the potential barrier, which is nominally at ram energy. The instrument requires pointing along the spacecraft’s ram velocity vector to make this measurement, thus requiring stabilization in pitch and yaw. During PLANE’s heritage flight, though the satellite’s attitude control system failed early in the mission, plasma data were collected during opportune times in which the instrument rotated into and out of the ram. Observations of plasma density and PSDs of high frequency plasma turbulence were recorded on several occasions. Additionally, a plasma source onboard the satellite was used to generate artificial plasma turbulence, and the PLANE data observed periodic structure presumably associated with the rotation of the spacecraft during these source firings. A brief comparison with other high frequency in situ plasma instruments is presented.

Forecasting trends of high-frequency KOSPI200 index data using learning classifiers

Recently many statistical learning techniques have been applied to the prediction of financial variables. The aim of this paper is to conduct a comprehensive study of the applications of statistical learning techniques to predict the trend of the return of high-frequency Korea composite stock price index (KOSPI) 200 index data using the information from the one-minute time series of spot index, futures index, and foreign exchange rate. Through experiments, it is observed that the spot index change is better predictable with high-frequency time series data and the futures index information significantly improves the prediction accuracy of the return trends of the spot index for high-frequency index data, while the information of exchange rate does not. Also, dimension reduction process before training helps to increase the accuracy and dramatically for some classifiers. In addition, the trained classifiers with which a virtual trading strategy is applied to, noticeable better profits can be achieved than just a buy-and-hold-like strategy.

High-frequency water quality time series in precipitation and streamflow: From fragmentary signals to scientific challenge

Eighteen months of 7-hourly analyses of rainfall and stream water chemistry are presented, spanning a wide range of chemical determinands and building on over 20years of weekly records for the moorland headwaters of the river Severn. The high-frequency time series data show that hydrochemical responses to major hydrological and biological drivers of short-term variability in rainfall and rivers are not captured by conventional low-frequency monitoring programmes. A wealth of flow related, flow independent, diurnal, seasonal and annual fluctuations indicate a cacophony of interactions within the catchment and stream. The complexity of the chemical dynamics is visually obvious, although there appears to be no clear way of translating this complexity into a simple algorithm. The work provides a proof of concept for the complex structure of catchment functioning revealed by extensive high-frequency measurements coupled with high analytical sensitivity and reproducibility. It provides new insights into hydrogeochemical functioning and a novel resource for catchment modelling.

Noninvasive Remote Sensing Techniques for Infrastructures Diagnostics

The present paper aims at analyzing the potentialities of noninvasive remote sensing techniques used for detecting the conservation status of infrastructures. The applied remote sensing techniques are ground-based microwave radar interferometer and InfraRed Thermography (IRT) to study a particular structure planned and made in the framework of the ISTIMES project (funded by the European Commission in the frame of a joint Call “ICT and Security” of the Seventh Framework Programme). To exploit the effectiveness of the high-resolution remote sensing techniques applied we will use the high-frequency thermal camera to measure the structures oscillations by high-frequency analysis and ground-based microwave radar interferometer to measure the dynamic displacement of several points belonging to a large structure. The paper describes the preliminary research results and discusses on the future applicability and techniques developments for integrating high-frequency time series data of the thermal imagery and ground-based microwave radar interferometer data.

Bivariate Long Term Fuzzy Time Series Forecasting of Dry Cargo Freight Rates

This paper proposes a bivariate long term fuzzy inference system for time series forecasting task in the field of freight market. Fuzzy time series methods are applied by many scholars, it is broadly accepted pattern recognition, forecasting tool. Previous studies mainly establish algorithms for high frequency time series data such as daily, monthly intervals. The proposed model performs similar techniques for long term annual base data, also extends the conventional method with multi-variate heuristic algorithm.Empirical work is accomplished on shipping freight rate data, life expectancy is used as a leading indicator in the bivariate fuzzy time series model.

IDENTIFYING NONLINEAR SERIAL DEPENDENCE IN VOLATILE, HIGH-FREQUENCY TIME SERIES AND ITS IMPLICATIONS FOR VOLATILITY MODELING

In this article, we show how tests of nonlinear serial dependence can be applied to high-frequency time series data that exhibit high volatility, strong mean reversion, and leptokurtotis. Portmanteau correlation, bicorrelation, and tricorrelation tests are used to detect nonlinear serial dependence in the data. Trimming is used to control for the presence of outliers in the data. The data that are employed are 161,786 half-hourly spot electricity price observations recorded over nearly a decade in the wholesale electricity market in New South Wales, Australia. Strong evidence of nonlinear serial dependence is found and its implications for time series modeling are discussed.

The properties and mechanism of long-term memory in nonparametric volatility

Recent empirical literature documents the presence of long-term memory in return volatility. But the mechanism of the existence of long-term memory is still unclear. In this paper, we investigate the origin and properties of long-term memory with nonparametric volatility, using high-frequency time series data of the Chinese Shanghai Composite Stock Price Index. We perform Detrended Fluctuation Analysis (DFA) on three different nonparametric volatility estimators with different sampling frequencies. For the same volatility series, the Hurst exponents reduce as the sampling time interval increases, but they are still larger than 1/2, which means that no matter how the interval changes, it still cannot change the existence of long memory. RRV presents a relatively stable property on long-term memory and is less influenced by sampling frequency. RV and RBV have some evolutionary trends depending on time intervals, which indicating that the jump component has no significant impact on the long-term memory property. This suggests that the presence of long-term memory in nonparametric volatility can be contributed to the integrated variance component. Considering the impact of microstructure noise, RBV and RRV still present long-term memory under various time intervals. We can infer that the presence of long-term memory in realized volatility is not affected by market microstructure noise. Our findings imply that the long-term memory phenomenon is an inherent characteristic of the data generating process, not a result of microstructure noise or volatility clustering.

On the Anomalous Behaviour of Scalar Flux–Variance Similarity Functions Within the Canopy Sub-layer of a Dense Alpine Forest

Within the canopy sub-layer (CSL), variability in scalar sources and sinks are known to affect flux–variance (FV) similarity relationships for water vapour (q) and carbon dioxide (C) concentrations, yet large-scale processes may continue to play a significant role. High frequency time series data for temperature (T), q and C, collected within the CSL of an uneven-aged mixed coniferous forest in Lavarone, Italy, are used to investigate these processes within the context of FV similarity. This dataset suggests that MOST scaling describes the FV similarity function of T even though the observations are collected in the CSL, consistent with other studies. However, the measured FV similarity functions for q and C appear to have higher values than their temperature counterpart. Two hypotheses are proposed to explain the measured anomalous behaviour in the FV similarity functions for q and C when referenced to T. Respired CO2 from the forest floor leads to large positive excursions in the C time series at the canopy top thereby contributing significantly to both C variance increase and C vertical flux decrease—both leading to an anomalous increase in the FV similarity function. For q, transport of dry air from the outer-layer significantly increases both the variance and the water vapour flux. However, the expected flux increase is much smaller than the variance increase so that the net effect remains an increase in the measured FV similarity function for water vapour above its T counterpart. The hypothesis here is that identifying these events in the temporal and/or in the frequency domain and filtering them from the C and q time series partially recovers a scalar flow field that appears to follow FV similarity theory scaling. Methods for identifying both types of events in the time and frequency domains and their subsequent effects on the FV similarity functions and corollary flow variables, such as the relative transport efficiencies, are also explored.

Competition in Intermediated Markets: Statistical Signatures and Critical Densities

In conventional economic theory, competition is an unmodelled process that is claimed to drive all economic actors to behave as if they were constrained optimisers. What is actually modelled in conventional economic theory is either a competitive equilibrium that is said to capture the result of the unmodelled competitive process or a game between two or, occasionally, among three agents. The common element in these approaches is the absence of any consideration of the effect of interaction among more than three economic actors at any one time. This is despite the natural presumption that such interaction is essential to any process of competition. The purpose of this paper is to demonstrate that the interaction excluded from conventional economic theory gives rise to the distributions of data observed in real markets. The well known fat-tailed (leptokurtic) distributions found in high frequency time series data for financial markets is shown to characterise high frequency retail market data as well. This and other statistical signatures of real markets are replicated in an abstract simulation model of markets in which intermediaries (brokers or jobbers) can function. It is shown that a high density of agents in the social network is necessary for intermediation to be viable and for the preponderance of demands to be satisfied. Moreover, the leptokurtic statistical signature characterises only markets satisfying that density condition.

Betting volume and market efficiency in Hong Kong race track betting

The Harville formula provides an analytical basis for the arbitrage pricing relationship between the odds on ‘win’ bets and the odds on ‘quinella’ bets in parimutuel betting markets. This study employs this relationship and high frequency time series data on the evolution of the betting odds from Hong Kong race tracks to test the hypothesis of betting market efficiency and to examine the role of the volume of belting. The non–stationary odds data are analysed using Johansen's maximum likelihood cointegration techniques to test the condition for market efficiency. The volume of betting is found to be a significant determinant of the degree of market efficiency across races but not within races.