Intraday Trends Research Articles

Prediction of Equities Basis Intraday Movement

Abstract: Data is a significant driver for several business decisions in this current era due to advent of big data and Stock markets are a great playground for deep learning and AI due to their high uncertainty. Although there are several tools in the market that aim to predict the future stock values with a prediction accuracy of almost 85%, none of them is fool proof and pave way and create scope for better accuracy. The current suite of tools predicts intraday buy /sell prices but are not capable of generating the pattern of day trade and stock swing, which when predicted, can generate 12X revenues for a full-time trader. The work focuses on the intraday trend and aims to predict the flow of Intraday trade, to gain larger profits and minimize losses in time in case of adverse conditions.

A novel feature engineering approach for high-frequency financial data

Feature engineering for high-frequency financial data based on constructing dynamic data subsets, defined by time intervals in which high-frequency trends occur, is proposed. These intervals are obtained through time series segmentation. This methodology allows us to extract and analyze variables by intraday trends as well as to feed artificial intelligence models to forecast response variables in future trends. Furthermore, to show how to use this feature engineering, this methodology is applied to estimate high-frequency volatility, duration and direction linked to future intraday trends, developing multiclass classification models based on the machine learning method extreme gradient boosting. Experimentation was conducted using high-frequency financial data from the Brazil Stock Exchange, corresponding to 206 trading days related to 20 listed assets from this financial market.

Intraday trend prediction of stock indices with machine learning approaches

In recent years, as research at the intersection of machine learning and finance has grown, predicting stock price movements has become a particularly intriguing issue. Current research focuses primarily on using historical data of the previous day to predict stock movements for the following day, whereas fewer studies use the trading day’s opening data to predict market movements for the current day. We predict intraday price movements of the SSE-50 (Shanghai Securities 50 Index) using stock market opening data as input. Specifically, decision tree, extreme gradient boosting (XGBoost), random forest, support vector machines (SVM), and long-short-term memory are developed to predict the movements of the SSE-50 index utilizing opening price data of various time intervals. We also design three trading strategies when different time frequencies of data are used. At the same time-frequency, the results demonstrate that SVM with Gaussian and linear kernels outperform others. The forecasting accuracy at 10-min frequency approaches 70%, which is close to the results at longer time intervals, indicating that intraday trend can be determined by opening price fluctuations and the first 10-min data contains sufficient information to predict the trend for the entire trading day. In addition, trading methods based on the forecast of daily, weekly, and monthly SSE-50 price movement outperform buy-and-hold strategies. Daily trading performs better than the other two strategies. The outcomes of this research can expand the use of machine learning in quantitative trading and enrich intraday trading techniques further.

Intraday Return Forecasts and High-Frequency Trading of Stock Index Futures: A Hybrid Wavelet-Deep Learning Approach

ABSTRACT We propose a novel hybrid wavelet-deep learning (DB-BLSTM) model to cope with the complex periodicity and nonlinearity issues in high-frequency data, which make the traditional linear time-series prediction models not applicable and result in weak predictability. The DB-BLSTM model we initiated in the paper can significantly outperform other deep learning models in predicting the intraday trends of Chinese stock index futures for both in-sample and out-of-sample tests. Trading strategies based on the DB-BLSTM models can achieve excellent excess returns and impressive return compensation relative to risks, and at the same time they can effectively control drawdown risk.

The Early Bird Catches the Intraday Trend

Large movements during the night tend to have a lasting influence on the next trading session. By following the resulting trends, investors can benefit from fluctuations in the returns of major equity indices during the trading session. This study extends recent research on late-day trading and shows that trends are best caught early during the day. Intraday patterns are revealed using machine learning techniques, and new evidence on the origin of trends is presented. The defensive nature of intraday trend strategies is illustrated by considering the business cycle, investigating correlation with government bonds, and simulating the impact on a multi-asset portfolio.

АНАЛІЗ ЗАСТОСУВАННЯ ІНСТРУМЕНТІВ РИНКОВОГО ПРОФІЛЮ ТА МЕТОДИКИ СКАЛЬПІНГА У ВНУТРІШНЬОДЕННІЙ ТОРГІВЛІ Ф’ЮЧЕРСНИМИ КОНТРАКТАМИ.

The article discusses the tools of the theory of market profile used to analyze the distribution of exchange volumes, as well as the principles of practical application of the analysis of the flow of orders and scalping elements. This paper proposes its own intraday trading strategy based on the assumption that the traders have open unprofitable positions on the market based on data from the report on changes in the open interest of the Chicago Mercantile Exchange and their identification using non-standard application of market profile and scalping tools, which makes it possible to adopt effective trading decisions. Using the tools for analyzing the volume of transactions of market auction participants - the levels of the value area, we identified the zones of possible concentration of open unprofitable positions, which in the future can be massively closed, creating a price impulse. Using scalping tools - Level I and Level II (the flow of transactions and market depth), changes in the nature of the dynamics of the flow of orders of participants in a market auction were analyzed. The strategy was tested on real stock data of a futures contract on the E-mini S&P 500 stock index by real-time testing. The test results confirm the assumption that there is a constant presence of open unprofitable positions in the market, since transactions based on the search for concentration zones of these positions were made against the direction of the price trend, which moves the market to identified zones. The massive closure of unprofitable positions, later confirmed by orders of bidders, turned out to be able to unfold the current market intraday trend, which makes it possible to make effective trading decisions and in the future increase trade capital. The use of exchange volume analysis tools in trading has proven to be very effective and can complement any intraday techniques, increasing profitability and lowering risk indicators in trading. The test results showed a critical effect of commission costs on intraday trading methods, which can significantly reduce the profitability indicator up to a loss-making strategy result

Dynamic Trade Informativeness

This paper develops a structural model to examine high-frequency price dynamics. The key innovation is to allow trades’ permanent price impact to be time-varying—dynamic trade informativeness. A distribution-free filtering technique pins the real-world data to the model. The filtered series significantly recover the efficient price innovation through the dynamics of trade informativeness, improve trades’ explanatory power for future returns, distinguish informativeness from trades’ aggressiveness, gauge informed investors’ patience, and capture systematic patterns around scheduled and unscheduled events, as well as general intraday trends. The framework contributes to the better utilization of high-frequency trading data.

Modeling the Spatial Patterns of Intra-Day Crime Trends

Several prior studies have found that despite theoretical expectations otherwise, facilities (such as on-premise alcohol outlets) have consistent effects on crime regardless of time of the day (Bernasco et al., 2017; Haberman & Ratcliffe, 2015). We explain these results by failure to account for the regular background wave of crime, which results from ubiquitous patterns of human routine activities. Using eight years of data on assaults and robberies in Seattle (WA), we demonstrate the regularity of the within-day crime wave for all areas of the city. Then using models to predict when a crime will most likely occur, we demonstrate how schools and on-premise alcohol outlets cause bumps in the background wave at particular times of the day, such as when school dismisses. But those bumps dissipate quite rapidly in space, and are relatively small compared to the amplitude of the regular background wave of crime. Although facilities have theoretical times in which they should have a greater influence on crime patterns, they are situated within a community of other human activity uses, making it difficult to uniquely identify their effects separately from other aspects of the built environment.

Traffic flow data compression considering burst components

Many recent applications of intelligent transportation systems require both real-time and network-wide traffic flow data as input. However, as the detection time and network size increase, the data volume may become very large in terms of both dimension and scale. To address this concern, various traffic flow data compression methods have been proposed, which archive the low-dimensional subspace rather than the original data. Many studies have shown the traffic flow data consist of different components, i.e. low-dimensional intra-day trend, Gaussian type fluctuation and burst components. Existing compression methods cannot compress the burst components well and provide very limited choices of compression ratio (CR). A better compression method should have the ability to archive all the dominant information in different components of traffic flow data. In this study, the authors compare the influence of different data reformatting, archive the bursts defined before in descending order with respect to the absolute value of the burst points and propose a flexible compression framework to balance between burst components and low-dimensional intra-day trend. Experimental results show that the proposed framework promotes the reconstruction accuracy significantly. Moreover, the proposed framework provides more flexible choices with respect to CR, which can benefit a variety of applications.

Grey Coupled Prediction Model for Traffic Flow with Panel Data Characteristics

This paper studies the grey coupled prediction problem of traffic data with panel data characteristics. Traffic flow data collected continuously at the same site typically has panel data characteristics. The longitudinal data (daily flow) is time-series data, which show an obvious intra-day trend and can be predicted using the autoregressive integrated moving average (ARIMA) model. The cross-sectional data is composed of observations at the same time intervals on different days and shows weekly seasonality and limited data characteristics; this data can be predicted using the rolling seasonal grey model (RSDGM(1,1)). The length of the rolling sequence is determined using matrix perturbation analysis. Then, a coupled model is established based on the ARIMA and RSDGM(1,1) models; the coupled prediction is achieved at the intersection of the time-series data and cross-sectional data, and the weights are determined using grey relational analysis. Finally, numerical experiments on 16 groups of cross-sectional data show that the RSDGM(1,1) model has good adaptability and stability and can effectively predict changes in traffic flow. The performance of the coupled model is also better than that of the benchmark model, the coupled model with equal weights and the Bayesian combination model.

An analysis of aerosol liquid water content and related impact factors in Pearl River Delta.

Aerosol liquid water content (ALWC) has an important effect on atmospheric visibility as well as heterogeneous chemical reactions. In this paper, we used the data size-resolved particle hygroscopic growth factor, and particle number size distribution (PNSD) obtained from H-TDMA and SMPS to compute ALWC at the Guangzhou Panyu site from the winter of 2014 and the spring of 2015. The corresponding results were relatively consistent with the trend for ALWCISO calculated from the ISORROPIA II thermodynamic equilibrium model based on the measurement of aerosol water-soluble ionic compositions obtained from MARGA, with a linear fit yielding an R2 value of 0.76. The fact that ALWCHTDMA was somewhat higher than ALWCISO at low RH values was at least partially attributable to the fact that effects resulting from organic matter hygroscopicity were not taken into account when computing ALWCISO. In sensitivity testing, ambient relative humidity, PNSD and particle hygroscopicity were all found to affect ALWC, in that order. Particles of different modes made different contributions to ALWC with the contributions of nuclear, Aitken, accumulation and coarse modes assessed at <1%, 3%, 85% and 12%, respectively, indicating that the contribution of accumulation mode particles to ALWC dominated among all the aerosol particle modes. During clean processes, decreases in relative humidity and PM2.5 both resulted in a decrease in ALWC. During the pollution processes, calm winds caused local particle accumulation, with ALWC increasing as RH increased. Intraday trends in ALWC and relative humidity were consistent, with minimum mean values observed in the afternoon due to low ambient relative humidity inhibiting an increase in ALWC. However, diurnal variation of aerosol hygroscopicity and ALWC tended to be somewhat anti-correlated, indicating that diurnal changes in aerosol hygroscopicity are not a primary factor resulting in ambient AWLC changes.

Time-dependent scaling patterns in high frequency financial data

We measure the influence of different time-scales on the dynamics of financial market data. This is obtained by decomposing financial time series into simple oscillations associated with distinct time-scales. We propose two new time-varying measures: 1) an amplitude scaling exponent and 2) an entropy-like measure. We apply these measures to intraday, 30-second sampled prices of various stock indices. Our results reveal intraday trends where different time-horizons contribute with variable relative amplitudes over the course of the trading day. Our findings indicate that the time series we analysed have a non-stationary multifractal nature with predominantly persistent behaviour at the middle of the trading session and anti-persistent behaviour at the open and close. We demonstrate that these deviations are statistically significant and robust.

Emissions reduction benefits of siting an offshore wind farm: A temporal and spatial analysis of Lake Michigan

Siting decisions of offshore wind farms influence the magnitude of emissions reduction benefits. This paper calculates electricity generation and emissions reduction of CO2, NOx, and SO2, and values these reductions to determine the impact of the siting location for a 300MW offshore wind farm in Lake Michigan. The most important patterns for emissions reduction were the monthly trends, where January, March, and December consistently had the highest electricity generation and emissions reduction benefits. Summer months such as July and August had the lowest emissions reduction benefits. The intra-day trends showed higher emissions reduction benefits during off-peak hours, due to a higher likelihood of coal units being the marginal generator. These diurnal differences were smaller in magnitude than the seasonal differences. Two benefit valuation scenarios were analyzed for a 20-year time period, one using marginal damages of pollution and another using market prices for pollution allowances. The first scenario resulted in emissions reduction benefits ranging from $1827/kW to $2690/kW ($2508/kW averaged) throughout the Lake Michigan region for the 20-year period (applying a 3% discount rate). This equates to approximately $33/MWh in all lake locations since the emissions reduction benefits are primarily a function of electricity generation. The market price scenario resulted in a much lower range of $820/kW to $1060/kW ($987/kW average or 39% of the pollution damage costs). In scenario 1, the major component of emissions reduction benefits was CO2 reduction (86% of benefits), and 83% of these CO2 benefits were from offsetting coal plant emissions. A sensitivity analysis on size and region of emission reduction location showed that the NOx and SO2 benefits vary significantly (unlike CO2 benefits), but this variation had minimal effects on the total emissions reduction benefits. In comparison with economic investment costs, the scenario 1 emissions reduction benefits equal 49% of the total investment cost (in 2014 $million) on average. Spatial maps and heat maps are generated to illustrate the spatial and temporal variations in the emissions reduction benefits.

A New Traffic Prediction Method based on Dynamic Tensor Completion

Traditional traffic prediction methods treat traffic data as one dimensional time series that can’t make full use of multi-mode correlation of traffic data, hence previous prediction models exist different levels of predictability and limits. To fully utilize the intrinsic multiple correlations of traffic data, in this paper a multi dimensional array-tensor model has proposed to encapsulate the traffic volume data. And a new traffic prediction method has been proposed, which includes rough estimation with intra-day trend and exact estimation with dynamic tensor completion (DTC) Experimental results demonstrate that the proposed prediction method is more accurate and reliable than traditional prediction methods.

Multi‐scale relation analysis of power law distribution and correlation in the Chinese stock market

PurposeThe purpose of this paper is to reveal the multi‐scale relation between power law distribution and correlation of stock returns and to figure out the determinants underlying capital markets.Design/methodology/approachThe multi‐scale relation between power law distribution and correlation is investigated by comparing the original series with the special series. The eliminating intraday trend series approach developed by Liuet al.is utilized to analyze the effects of power law decay change on correlation properties, and shuffling series originated by Viswanathanet al.for the impacts of special type of correlation on power‐law distribution.FindingsIt is found that the accelerating decay of power law has an insignificant effect on correlation properties of returns and the empirical results indicate that time scale may also be an important factor maintaining power law property of returns besides correlation. When time scale is under critical point, the effects of correlation are crucial, and the correlation of nonlinear long‐range presents the strongest influence. However, for time scale beyond critical point, the impact of correlation begins to diminish or even finally disappear and then the power law property shows complete dependence on time scale.Research limitations/implicationsThe 5‐min high frequency data of the Shanghai market as the empirical benchmark is insufficient to depict the relation over the entire time scale in the Chinese stock market.Practical implicationsThe paper identifies the determinants of market dynamics to apply them to risk management through analysis of multi‐scale relations, and supports endeavors to introduce time parameter into further risk measures and control.Originality/valueThe paper provides the empirical evidence that time scale is one of the key determinants of market dynamics by analyzing the multi‐scale relation between power law distribution and correlation.

The retrieval of intra-day trend and its influence on traffic prediction

In this paper, we discuss three problems that occur within short-term traffic prediction when the information from only a single point loop detector is used. First, we analyze the retrieval of intra-day trend for traffic flow series and determine whether this retrieval process improves traffic prediction. We compare different highway traffic prediction models that use either the original traffic flow series or the residual time series with the intra-day trend removed. Test results indicate that the prediction performance MAY be significantly improved in the latter scenario. Second, we address two other related questions: the influence of missing data and traffic breakdown prediction. We show that the Probabilistic Principal Component Analysis (PPCA) method, which also utilizes the intra-day trend of traffic flow series, can be a useful tool in imputing the missing data. It can simultaneously ensure that the prediction error remains at an acceptable level, especially when the missing ratio is relatively low. We also show that almost all the known predictors have hidden assumptions of smoothness and, thus, cannot predict the burst points that deviate too far from the intra-day trend. As a result, traffic breakdown points can only be identified but not predicted.

On-line VWAP Trading Strategies

VWAP stands for volume-weighted average price during a certain trading period, and a VWAP trade refers a trade that uses VWAP as a benchmark. This article provides on-line execution strategies for a VWAP trade. We first propose a simulation-based statistical price-volume model that enables the VWAP to be reformulated as a combination of two Brownian motions. Then we introduce the dynamic programming algorithm as the comparison of our trading strategies. Since the dynamic programming algorithm is model dependent and computationally intensive, we present some simple alternative strategies for VWAP trading. Among these, we first propose a modified cross-boundary strategy which can be shown as an asymptotic approximation of the dynamic programming strategy. Next, we introduce a relative rank strategy that ignores the actual stock price and considers only the stock price's relative rank. Our simulation results show that the (modified) cross-boundary strategy is better for stock prices with negative drift, whereas the relative rank strategy performs well for stock prices with positive drift. As a result, when a trader faces a long trading horizon with multi-periods and variate drifts, a hybrid algorithm based on the intra-day drift trend is proposed. Finally, to demonstrate the robustness of the trading strategies relative to the price-volume model, and evaluate the effect of VWAP trading, we present an empirical study of the trading strategies on the top 20 liquidity stocks on the Taiwan Stock Exchange Corporation.

Lead-lag associations between option trading and cash market volatility

The paper examines the relationships between the OEX trading volume and interday and intraday trends in the corresponding S&P100 cash index for the 11-year interval, February 1984 to April 1995. The evidence is consistent with the notion that option trading has a stabilizing impact on cash markets. Specifically, interday and intraday volatility is found to decline following increased option trading activity. This pattern is documented for total option trading activity, speculative trading activity, and across option trading activities broken down by put and call volume. Patterns of causality from cash markets to option trading are not as evident