Ultra-high Frequency Data Research Articles

Pattern Recognition of Stock Returns in the Very Short Run Leveraging High-Frequency Financial Data

This paper uses high frequency stock trading data from 2007 to 2014 to study patterns of stock returns in the very short run. The paper verifies stylized facts of stock prices at low frequency with the exploration of high-frequency data and uses the concept of relative realized volatility (RRV) to measure volatility to understand the market uncertainty intraday. By providing a large number of empirical data facts, this paper advocates the use of ultra-high frequency data to study instantaneous real volatility, and demonstrates that long-term market volatility and the relationship between short and long term volatility can be implied by a simple RRV mean regression model.

The Heterogeneity of Investors Based on Multi-fractal Features with Ultra-High Frequency Data

The Heterogeneity of Investors Based on Multi-fractal Features with Ultra-High Frequency Data

UHF-HF Integrated RFID Sensor: A Novel Information Interface for Industrial Integration

The itemized product identification, sensing and data integration are critical building blocks for the efficiency, flexibility, and potential intelligence of the Industrial Internet of Things (IIoT). The product sensing and data integration solutions usually suffer from the sophisticated sensor devices and their inflexible integration with the industrial processes, which may restrict the performance of industrial systems. To maintain an information link for convenient product data integration, this investigation presents a surface mountable dual-mode Radio Frequency Identification (RFID) sensor as a wirelessly powered information interface for simultaneous product identification, sensing, and data integration. The focus of this investigation includes: (1) the theory and hardware/software scheme for sensor-augmented RFID integrating Ultra-High Frequency (UHF) and High Frequency (HF) with Commercial-off-the-Shelf (COTS) devices; (2) integration of the UHF and HF energy and data interfaces to support both UHF and HF powering and sensing; (3) a data integration method supporting sensor data transmission using common UHF and HF readers without extra hardware or software cost; (4) proposal of a model for industrial integration with the proposed dual-mode RFID sensor. The experimental studies and exemplar application demonstrate the feasibility of the presented techniques for industrial integration.

A Novel Method for Pattern Recognition of GIS Partial Discharge via Multi-Information Ensemble Learning.

Partial discharge (PD) is the main feature that effectively reflects the internal insulation defects of gas-insulated switchgear (GIS). It is of great significance to diagnose the types of insulation faults by recognizing PD to ensure the normal operation of GIS. However, the traditional diagnosis method based on single feature information analysis has a low recognition accuracy of PD, and there are great differences in the diagnosis effect of various insulation defects. To make the most of the rich insulation state information contained in PD, we propose a novel multi-information ensemble learning for PD pattern recognition. First, the ultra-high frequency and ultrasonic data of PD under four typical defects of GIS are obtained through experiment. Then the deep residual convolution neural network is used to automatically extract discriminative features. Finally, multi-information ensemble learning is used to classify PD types at the decision level, which can complement the shortcomings of the independent recognition of the two types of feature information and has higher accuracy and reliability. Experiments show that the accuracy of the proposed method can reach 97.500%, which greatly improves the diagnosis accuracy of various insulation defects.

A Comprehensive Review of Signal Processing and Machine Learning Technologies for UHF PD Detection and Diagnosis (I): Preprocessing and Localization Approaches

Partial discharge (PD) detection and diagnosis based on the ultra-high frequency (UHF) signals is one of the most widely adopted methods to evaluate the internal insulation status of high voltage equipment. Benefit from the rapid development of computing hardware and data processing algorithms, the intelligent PD fault diagnosis method based on the UHF data has made considerable progress in the past two decades. This two-part paper aims to give a comprehensive review about the application of signal processing and machine learning technologies in UHF PD detection and diagnosis. These technologies are divided into three categories according to their respective purpose, which are the preprocessing technology, source localization technology and pattern recognition technology. As the first one of the two-part review, we focus on the preprocessing and localization approaches in this paper. Specifically, for the preprocessing topic, the methods for signal denoising, multi-source separation, and pulse segmentation are included. While for the localization topic, the time difference of arrival (TDOA) method, direction of arrival (DOA) method, received signal strength indicator (RSSI) method, and other latest methods are reviewed. For each topic, the basic ideas, recent research progresses, advantages and limitations are discussed in detail. Before the conclusion, we also make a discussion about the application effects of the above technologies and prospect some future directions accordingly. In the second paper, the pattern recognition problems based on the UHF PD data will be concentrated, especially the application of deep learning algorithms.

Empirical scaling relations of market event rates in foreign currency market

The mechanism of order book dynamic is studied by using ultra high frequency data in terms of three market events: injection, cancellation, and transaction. We analyzed the empirical decision-making process of market participants focusing on the event rates conditional on the depth from the mid-price. We observed that both injection and cancellation rates depend on the market depth with exponential decay near the mid-price, which is different from the conventional assumption of the homogeneous Poisson process for the order book formation. We also found scaling relations between injection, cancellation, transaction, and diffusion of the mid-price, highlighting the correlation between these market events. We finally discussed a theoretical model based on our findings to reproduce the empirical order-book profiles.

Forecasting market risk using ultra-high-frequency data and scaling laws

This paper develops a new multiple time scale-based empirical framework for market risk estimates and forecasts. Ultra-high frequency data are used in the empirical analysis to estimate the parameters of empirical scaling laws which gives a better understanding of the dynamic nature of the market. A comparison of the new approach with the popular Value-at-Risk and expected tail loss measures with respect to their risk forecasts during the crisis period in 2008 is presented. The empirical results show the outperformance of the new scaling law method which turns out to be more accurate and flexible due to the scale invariance. The scaling law method promotes the use of massive real data in developing risk measurement and forecasting models.

A high-frequency analysis of price resolution and pricing barriers in equities on the adoption of a new currency

ABSTRACTWe use ultra high frequency (trade by trade) data to demonstrate that equity price clustering and pricing predictability around psychologically important prices in Greece switches away from drachma-focused with the introduction of the euro, but does not immediately switch to euro-clustering. The change in trader price focus around the euro introduction addresses an open debate in the clustering literature on whether the presence of clustering is a bias related to the current prices or anchoring to past prices. Our findings of a decline in drachma clustering, but lack of switch to euro effects supports the case for clustering being a trading feature that is slow to transfer to new pricing regimes. A key advantage of the ultra high frequency dataset is we are also able to demonstrate the presence of psychological pricing barriers related to each currency that are not detectable in daily data.

Bayesian Inference via Filtering Equations for Ultrahigh Frequency Data (I): Model and Estimation

We propose a general partially observed framework of Markov processes with marked point process observations for ultrahigh frequency (UHF) data. The model fits well the stylized facts of UHF data in both macro- and micro-movements, subsumes important existing models, and incorporates the influence of other observable economic or market factors. We develop the corresponding Bayes estimation via a filtering equation to quantify parameter uncertainty. Namely, we derive the normalized filtering equation to characterize the evolution of the posterior distribution, present a weak convergence theorem, and construct consistent, easily parallelizable, recursive algorithms to calculate the joint posteriors and the Bayes estimates for streaming UHF data. Moreover, a sufficient condition for the consistency of the Bayes estimators is provided. The general estimation theory is illustrated by four specific models built for U.S. Treasury notes transactions data from GovPX. We show that in this market, both information-b...

Bayesian Inference via Filtering Equations for Ultrahigh Frequency Data (II): Model Selection

For the general partially observed framework of Markov processes with marked point process observations proposed in [G. X. Hu, D. R. Kuipers, and Y. Zeng, SIAM/ASA J. Uncertain. Quantif., 6 (2018),...

Feature Selection for Partial Discharge Severity Assessment in Gas-Insulated Switchgear Based on Minimum Redundancy and Maximum Relevance

Scientific evaluation of partial discharge (PD) severity in gas-insulation switchgear (GIS) can assist in mastering the insulation condition of in-service GIS. Limited theoretical research on the laws of PD deterioration leads to a finite number of evaluation features extracted and subjective features selected for PD severity assessment. Therefore, this study proposes a minimum-redundancy maximum-relevance (mRMR) algorithm-based feature optimization selection method to realize the scientific and reasonable choice of PD severity features. PD ultra-high frequency data of varying severities are produced by simulating four typical insulation defects in GIS, which are then collected in the lab. A 16-dimension feature set describing PD original characteristics is abstracted in phase-resolved partial discharge (PRPD) mode, and the more informative evaluation feature set characterizing PD severity is further excavated by the mRMR method. Finally, a support vector machine (SVM) algorithm is employed as the classifier for intelligent evaluation to compare the evaluation effects of PD severity between the feature set selected by mRMR and the feature set is composed of discharge times, amplitude value, and time intervals obtained traditionally based on discharge change theory. The proposed comparison test showed the effectiveness of the mRMR method in informative feature selection and the accuracy of PD severity assessment for all defined defects.

Inference for Time-Varying Lead-Lag Relationships from Ultra High Frequency Data

A new approach for modeling lead-lag relationships in high frequency financial markets is proposed. The model is accommodated to non-synchronous trading and market microstructure noise as well as intraday variations of lead-lag relationships, which are essential for empirical applications. A simple statistical methodology for analyzing the proposed model is presented as well. The methodology is illustrated by an empirical study to detect lead-lag relationships between the S&P 500 index and its two derivative products.

Stylized facts and trends of high frequency data in financial markets

High frequency data has given a paradigm shift to financial markets. High Frequency data is intra-daily data on financial variables taken at a finer time scale (also termed as tick size) and are often irregularly spaced over time. This data has an ability to define financial market micro-structure and make rational real-time decisions. There are diverse form of low frequency data like monthly, daily or yearly data but high frequency data, which means intra-day, intra-hour, intra-minutes or intra-seconds data have their own characteristics and special statistical properties, therefore also called as ultra high frequency data. Existing trends in high frequency literature revealed that there is huge prospective of this data for market researchers, traders, market practitioner, market regulators and investors in field of financial re-engineering, risk management systems, volatility modelling and forecasting.

The Dynamics of Ex-Ante High-Frequency Liquidity: An Empirical Analysis

The ex-ante liquidity embedded in an open Limit Order Book (LOB) and its dynamics have been one of the most important issues in financial research and evolves with the development of financial infrastructure. Using the tick-by-tick data and the reconstructed open LOB data from the Xetra trading system, we investigate the impact of trade duration, quote duration and other exogenous variables on ex-ante liquidity embedded in an open LOB. By taking into account Ultra High Frequency (UHF) data, our modeling involves decomposing the joint distribution of the ex-ante liquidity measure into simple and interpretable distributions. The decomposed factors are Activity, Direction and Size. Our results suggest that trade durations and quote durations do influence the ex-ante liquidity changes. Short-run variables, such as spread change and volume, also predict the probability of liquidity changes. The long-term variable trade imbalance is less informative.

Stylized facts of price gaps in limit order books

Price gap, defined as the logarithmic price difference between the first two occupied price levels on the same side of a limit order book (LOB), is a key determinant of market depth, which is one of the dimensions of liquidity. However, the properties of price gaps have not been thoroughly studied due to the less availability of ultrahigh frequency data. In the paper, we rebuild the LOB dynamics based on the order flow data of 26 A-share stocks traded on the Shenzhen Stock Exchange in 2003. Three key empirical statistical properties of price gaps are investigated. We find that the distribution of price gaps has a power-law tail for all stocks with an average tail exponent close to 3.2. Applying modern statistical methods, we confirm that the gap time series are long-range correlated and possess multifractal nature. These three features appear to be different in the measures across stocks, but they are similar for the buy and sell LOBs within each stock. Furthermore, we also unveil buy–sell asymmetry phenomena in the properties of price gaps on the buy and sell sides of the LOBs for individual stocks. These findings deepen our understanding of the dynamics of liquidity of common stocks and can be used to calibrate agent-based computational financial models.

Ultrahigh Frequency Data Liquidity Duration Estimation: A Case Study of Chinese A Shares

Liquidity has always been a hot spot for researchers of financial market microstructures. Analysis of liquidity is of great significance for investors and market regulators. Ultrahigh frequency data records the whole dynamic change of the trading process, so it has advantages in depicting the market microstructure. This study analyzes Asian emerging market equities liquidity using ultrahigh frequency data. We used various forms of WACD models and let trading duration be indicators of liquidity. Through the residual test, we were able to select the best model to describe the overall liquidity.

A microstructure analysis of the carbon finance market

The European Union Emissions Trading Scheme is the key policy instrument of the European Commission's Climate Change Program aimed at reducing greenhouse gas emissions to 8% below 1990 levels by 2012. The key asset traded under the scheme is the European Union allowance (EUA). This article examines ultra high frequency data to assess the extent of the development in the futures market of the EU Emissions Trading Scheme. Our results indicate significant developments consistent with sequential information arrival. They also indicate a negative contemporaneous relationship between volume and volatility for all contracts. The implication is that liquidity traders dominate any role played by informed traders. Incorporating the duration between trades in our analysis has significant impact suggesting that any empirical investigation of the intra-day volume–volatility relationship needs to actively account for the impact of time elapse between trades.

On the Intraday Relation between the VIX and Its Futures

The Chicago Board Options Exchange (CBOE) introduced the CBOE Volatility Index (VIX) in 1993. The index has come to act as the benchmark for stock market volatility and, more generally, investor sentiment. The VIX has proven to be very useful in forecasting the future market direction especially during high volatility periods. In order to expedite trading in volatility, as well as increase hedging opportunities, the CBOE introduced futures on the VIX (henceforth referred to as VXF) on March 26, 2004.We study the intraday dynamics of the VIX and VXF for the period January 2, 2008 to December 31, 2012. Applying a Vector Autoregression (VAR) model on daily data, we observe some evidence of causality from the VXF to the VIX. However, estimating a VAR using our ultra-high frequency data, we find strong evidence for bi-directional Granger causality between the VIX and the VXF. Overall, this effect appears to be stronger from the VXF to the VIX than the other way around. Impulse response functions and variance decompositions analysis further confirm the dominance of the VXF. Lastly, we show that the causality from the VXF to the VIX has been increasing over our sample period, whereas the reverse causality has been decreasing. This finding suggests that the VIX futures have become increasingly more important in the pricing of volatility. We further document that the VIX futures dominate the VIX more on days with negative returns, and on days with high values of the VIX, suggesting that on those days investors use VIX futures to hedge their positions rather than trading in the S&P 500 index options.

Bayesian Modelling of Integer Data Using the Generalised Poisson Difference Distribution

Integer-valued random variables arising from the difference of two discrete variables can be seen frequently in various applications. In this paper, we obtain the distribution and derive the properties of the difference of two generalised Poisson variables with unequal parameters. This distribution is adopted to model a set of ultra high frequency (UHF) data relating to FTSE100 index futures using covariates. The unique characteristics of UHF data have introduced new theoretical and computational challenges to both statistical and financial studies. Such data consist of discrete-valued observations and unequally spaced time intervals. We also extend the model to its zero inflated version in order to capture the excess of zeros in the given data set. The analysis is carried out in a Bayesian framework using Markov Chain Monte Carlo methods. Various model diagnostics and model comparisons were undertaken which showed that index changes were explained well by the fitted model.

Inference for a Nonstationary Self-Exciting Point Process with an Application in Ultra-High Frequency Financial Data Modeling

Self-exciting point processes (SEPPs), or Hawkes processes, have found applications in a wide range of fields, such as epidemiology, seismology, neuroscience, engineering, and more recently financial econometrics and social interactions. In the traditional SEPP models, the baseline intensity is assumed to be a constant. This has restricted the application of SEPPs to situations where there is clearly a self-exciting phenomenon, but a constant baseline intensity is inappropriate. In this paper, to model point processes with varying baseline intensity, we introduce SEPP models with time-varying background intensities (SEPPVB, for short). We show that SEPPVB models are competitive with autoregressive conditional SEPP models (Engle and Russell 1998) for modeling ultra-high frequency data. We also develop asymptotic theory for maximum likelihood estimation based inference of parametric SEPP models, including SEPPVB. We illustrate applications to ultra-high frequency financial data analysis, and we compare performance with the autoregressive conditional duration models.