Min Of Data Research Articles

Automatic detection and taxonomic identification of dolphin vocalisations using convolutional neural networks for passive acoustic monitoring

A novel framework for acoustic detection and species identification is proposed to aid passive acoustic monitoring studies on the endangered Indian Ocean humpback dolphin (Sousa plumbea) in South African waters. Convolutional Neural Networks (CNNs) were used for both detection and identification of dolphin vocalisations tasks, and performance was evaluated using custom and pre-trained architectures (transfer learning). In total, 723 min of acoustic data were annotated for the presence of whistles, burst pulses and echolocation clicks produced by Delphinus delphis (~45.6%), Tursiops aduncus (~39%), Sousa plumbea (~14.4%), Orcinus orca (~1%). The best performing models for detecting dolphin presence and species identification used segments (spectral windows) of two second lengths and were trained using images with 70 and 90 dpi, respectively. The best detection model was built using a customised architecture and achieved an accuracy of 84.4% for all dolphin vocalisations on the test set, and 89.5% for vocalisations with a high signal to noise ratio. The best identification model was also built using the customised architecture and correctly identified S. plumbea (96.9%), T. aduncus (100%), and D. delphis (78%) encounters in the testing dataset. The developed framework was designed based on the knowledge of complex dolphin sounds and it may assists in finding suitable CNN hyper-parameters for other species or populations. Our study contributes towards the development of an open-source tool to assist long-term studies of endangered species, living in highly diverse habitats, using passive acoustic monitoring.

Single-lead arrhythmia detection through machine learning: cross-sectional evaluation of a novel algorithm using real-world data

BackgroundComputer-assisted interpretation of single-lead ECG is the preliminary method for clinicians to flag and further evaluate an arrhythmia of clinical importance for acutely ill patients. Critical scrutiny of novel detection...

Statistical properties of cerebral near infrared and intracranial pressure-based cerebrovascular reactivity metrics in moderate and severe neural injury: a machine learning and time-series analysis

BackgroundCerebrovascular reactivity has been identified as a key contributor to secondary injury following traumatic brain injury (TBI). Prevalent intracranial pressure (ICP) based indices of cerebrovascular reactivity are limited by their invasive nature and poor spatial resolution. Fortunately, interest has been building around near infrared spectroscopy (NIRS) based measures of cerebrovascular reactivity that utilize regional cerebral oxygen saturation (rSO2) as a surrogate for pulsatile cerebral blood volume (CBV). In this study, the relationship between ICP- and rSO2-based indices of cerebrovascular reactivity, in a cohort of critically ill TBI patients, is explored using classical machine learning clustering techniques and multivariate time-series analysis.MethodsHigh-resolution physiologic data were collected in a cohort of adult moderate to severe TBI patients at a single quaternary care site. From this data both ICP- and rSO2-based indices of cerebrovascular reactivity were derived. Utilizing agglomerative hierarchical clustering and principal component analysis, the relationship between these indices in higher dimensional physiologic space was examined. Additionally, using vector autoregressive modeling, the response of change in ICP and rSO2 (ΔICP and ΔrSO2, respectively) to an impulse in change in arterial blood pressure (ΔABP) was also examined for similarities.ResultsA total of 83 patients with 428,775 min of unique and complete physiologic data were obtained. Through agglomerative hierarchical clustering and principal component analysis, there was higher order clustering between rSO2- and ICP-based indices, separate from other physiologic parameters. Additionally, modeled responses of ΔICP and ΔrSO2 to impulses in ΔABP were similar, indicating that ΔrSO2 may be a valid surrogate for pulsatile CBV.ConclusionsrSO2- and ICP-based indices of cerebrovascular reactivity relate to one another in higher dimensional physiologic space. ΔICP and ΔrSO2 behave similar in modeled responses to impulses in ΔABP. This work strengthens the body of evidence supporting the similarities between ICP-based and rSO2-based indices of cerebrovascular reactivity and opens the door to cerebrovascular reactivity monitoring in settings where invasive ICP monitoring is not feasible.

Personalized brain MRI revealed distinct functional and anatomical disruptions in Creutzfeldt-Jakob disease and Alzheimer's disease.

Creutzfeldt-Jakob disease (CJD) is a lethal neurodegenerative disorder, which leads to a rapidly progressive dementia. This study aimed to examine the cortical alterations in CJD, changes in these brain characteristics over time, and the differences between CJD and Alzheimer's disease (AD) that show similar clinical manifestations. To obtain reliable, subject-specific functional measures, we acquired 24 min of resting-state fMRI data from each subject. We applied an individual-based approach to characterize the functional brain organization of 10 patients with CJD, 8 matched patients with AD, and 8 normal controls. We measured cortical atrophy as well as disruption in resting-state functional connectivity (rsFC) and then investigated longitudinal brain changes in a subset of CJD patients. CJD was associated with widespread cortical thinning and weakened rsFC. Compared with AD, CJD showed distinct atrophy patterns and greater disruptions in rsFC. Moreover, the longitudinal data demonstrated that the progressive cortical thinning and disruption in rsFC mainly affected the association rather than the primary cortex in CJD. CJD shows unique anatomical and functional disruptions in the cerebral cortex, distinct from AD. Rapid progression of CJD affects both the cortical thickness and rsFC in the association cortex.

Wind Dependencies of Deep Cycle Turbulence in the Equatorial Cold Tongues

Abstract Several years of moored turbulence measurements from χpods at three sites in the equatorial cold tongues of Atlantic and Pacific Oceans yield new insights into proxy estimates of turbulence that specifically target the cold tongues. They also reveal previously unknown wind dependencies of diurnally varying turbulence in the near-critical stratified shear layers beneath the mixed layer and above the core of the Equatorial Undercurrent that we have come to understand as deep cycle (DC) turbulence. Isolated by the mixed layer above, the DC layer is only indirectly linked to surface forcing. Yet, it varies diurnally in concert with daily changes in heating/cooling. Diurnal composites computed from 10-min averaged data at fixed χpod depths show that transitions from daytime to nighttime mixing regimes are increasingly delayed with weakening wind stress τ. These transitions are also delayed with respect to depth such that they follow a descent rate of roughly 6 m h−1, independent of τ. We hypothesize that this wind-dependent delay is a direct result of wind-dependent diurnal warm layer deepening, which acts as the trigger to DC layer instability by bringing shear from the surface downward but at rates much slower than 6 m h−1. This delay in initiation of DC layer instability contributes to a reduction in daily averaged values of turbulence dissipation. Both the absence of descending turbulence in the sheared DC layer prior to arrival of the diurnal warm layer shear and the magnitude of the subsequent descent rate after arrival are roughly predicted by laboratory experiments on entrainment in stratified shear flows. Significance Statement Only recently have long time series measurements of ocean turbulence been available anywhere. Important sites for these measurements are the equatorial cold tongues where the nature of upper-ocean turbulence differs from that in most of the world’s oceans and where heat uptake from the atmosphere is concentrated. Critical to heat transported downward from the mixed layer is the diurnally varying deep cycle of turbulence below the mixed layer and above the core of the Equatorial Undercurrent. Even though this layer does not directly contact the surface, here we show the influence of the surface winds on both the magnitude of the deep cycle turbulence and the timing of its descent into the depths below.

Inflation and the war in Ukraine: Evidence using impulse response functions on economic indicators and Twitter sentiment

In this paper, we propose the use of social media information as a real-time decision-making tool for significant war events, using the war in Ukraine as a case study. We proxy the public’s perception of the progression of events using sentiment analysis on 42 million tweets and calculate impulse response functions on 5-min data for 15 economic and financial indicators. European indices (currencies and markets) experience an immediate negative response to conflict escalation “shocks”, while crude oil registers a delayed negative response. US stock markets seem unaffected, while the US Dollar responds positively to negative events of the war. Our findings suggest that user generated content can be used as a decision-making tool when important war events unfold. This approach can monitor the public’s perception of such events as well as capture their potential economic impact, which carries increased importance in times of increasing prices.

An Analytical Study on Soil Water Index (SWI), Landslide Prediction and Other Related Factors Using XRAIN Data during the July 2018 Heavy Rain Disasters in Hiroshima, Japan

The rainfall-induced landslide disasters in July 2018 in Southwestern Japan yet again exemplified the severity of slope failure-related damage and the need for improvement of early warning systems. The Japanese Meteorological Agency (JMA) uses a method based on a threshold value of soil water index (SWI), a conceptual measurement that represents saturation of slope soil. The current SWI early warning system uses 60-min rainfall data on a 5-km2 mesh and does not take into consideration other landslide conditioning factors such as slope angle and geology. This study calculates SWI values during the July 2018 disasters in Kure City (Hiroshima Prefecture) using 1-min XRAIN rainfall data in a 250-m mesh to investigate the relationship between SWI and landslide occurrence. It was found that the SWI threshold of 124 mm used in the JMA early warning system for the area was surpassed in all cells. A new SWI threshold calculation method taking slope angle and geology into consideration and produced with machine learning is proposed, comprising power lines for different geological units at a two-dimensional graph where points located above the threshold line represent landslide risk. It is judged that this method would provide a more accurate early warning system for landslide disasters.

Trajectory-based RFI subtraction and calibration for radio interferometry

ABSTRACT Radio interferometry calibration and radio frequency interference (RFI) removal are usually done separately. Here we show that jointly modelling the antenna gains and RFI has significant benefits when the RFI follows precise trajectories, such as for satellites. One surprising benefit is improved calibration solutions, by leveraging the RFI signal itself. We present trajectory-based RFI subtraction and calibration (tabascal), a new algorithm that jointly models the RFI and calibration parameters in visibilities. We test tabascal on simulated MeerKAT calibration observations contaminated by satellite-based RFI. We obtain gain estimates that are both unbiased and up to an order of magnitude better constrained compared to uncontaminated data. When combined with an ad hoc RFI subtraction scheme, tabascal solutions can be further applied to an adjacent target observation: 5 min of calibration data results in an image with about a third the noise achieved when using flagging alone. The recovered flux distribution of RFI subtracted data was on par with uncontaminated data. In contrast, RFI flagging alone resulted in a higher detection threshold and consistent underestimation of source fluxes. For a mean RFI amplitude of 17 Jy, using RFI subtraction leads to less than 1 per cent loss of data compared to 75 per cent data loss from an ideal 3σ flagging algorithm, a very significant increase in data available for science analysis. Although we have examined the case of satellite RFI, tabascal should work for any RFI moving on parametrizable trajectories, relative to the phase centre, such as planes and/or objects fixed to the ground.

Development of an Extreme Wind-Driven Rain Climatology for the Southeastern United States Using 1-Min Rainfall and Peak Wind Speed Data

Abstract A method is presented to obtain the climatology of extreme wind speeds coincident with the occurrence of rain. The simultaneous occurrence of wind and rain can force water through building wall components such as windows, resulting in building damage and insured loss. To quantify this hazard, extreme value distributions are fit to peak 3-s wind speed data recorded during 1-min intervals with specific reported rain intensities. This improves upon previous attempts to quantify the wind-driven rain hazard that computed wind speed and rainfall-intensity probabilities independently and used hourly data that cannot assure the simultaneous occurrence of peak wind that represents only a several-second interval within the hour and rain that is accumulated over the entire hour. The method is applied across the southeastern United States, where the wind-driven rain hazard is most pronounced. For the lowest rainfall intensities, the computed wind speed extremes agree with published values that ignore rainfall occurrence. Such correspondence is desirable for aligning the rain-intensity-dependent wind speed return periods with established extreme wind statistics. Maximum 50-yr return-period wind speeds in conjunction with rainfall intensities ≥0.254 mm min−1 exceed 45 m s−1 in a swath from Oklahoma to the Gulf Coast and at stations along the immediate Atlantic coast. For rainfall intensities >2.54 mm min−1 maximum, 50-yr return-period wind speeds decrease to 35 m s−1 but occur over a similar area. The methodology is also applied to stations outside the Southeast to demonstrate its applicability for incorporating the wind-driven rain hazard in U.S. building standards. Significance Statement Rainfall driven horizontally by strong winds can penetrate building components and cladding. If unmanaged, this can directly damage the building and its contents and become a substantial component of insured losses to buildings. A climatology of wind-driven rain is developed from recently available 1-min weather observations that better represent the joint occurrence of the extremes that define wind-driven rain occurrence than hourly data. This work is a first implementation of 1-min data into extreme-value statistical models, providing a basis for including wind-driven rain in United States building codes. This inclusion would be most significant in the hurricane-prone regions of the southeastern United States. The omission of wind-driven rain in U.S. building codes contrasts to its inclusion in Europe and Canada.

Teachers’ perceptions of technical affordances in early visual arts education

ABSTRACT This study aimed at exploring kindergarten teachers’ perceptions and usage of digital devices in their classrooms. A multiple case study (n = 4) was conducted to investigate how in-service experienced (n = 22) and novice (n = 24) kindergarten teachers interpreted the functional significance of digital devices, including the needs and challenges faced when applying digital devices in visual arts activities in their classrooms. A non-randomised convenience sampling strategy was used to recruit four local kindergartens in Hong Kong, whose teachers were invited to participate in focus group interviews. In total, 540 min of interview data were collected. We examined the in-service teachers’ perceptions of the affordances of digital devices in terms of (1) transparency, (2) challenge and (3) accessibility. Based on the teachers’ responses in the interviews, the findings indicated issues related to the use of digital devices in terms of these three areas here. The informants shared how their understandings of film language limited their intentions and usage patterns (including frequency of use) with respect to digital videos. Nevertheless, they admitted that digital videos were important tools for facilitating children’s learning and communication with their parents. This study reflects the current situation of technology education in Hong Kong, including how teacher educators perceive the usage and roles of digital videos in early arts education.

Intra-hour PV power forecasting based on sky imagery

This paper introduces an upgraded version of the PV2-state model [Paulescu et al. Renew Energy 195 (2022) 322] for intra-hour photovoltaic (PV) power forecasting. The model incorporates real-time adjustments to the estimated clear-sky PV power or the transmittance of clouds, considering the Sun coverage by clouds. A physics-based approach for processing cloud field information from an all-sky imager is proposed to intra-hour forecast SSN, a binary quantifier stating whether the Sun is shining (SSN = 1) or not (SSN = 0). The model performance with the new procedure is investigated from three perspectives: forecast accuracy, forecast precision and the response to the variability in the state-of-the-sky. The study was conducted with high-quality 1-min data collected from a micro-PV plant, on a sample of days with scattered clouds. PV2-state demonstrates notable performance particularly under challenging conditions, where models typically struggle to perform well. In terms of skill score, the overall model performance ranges between 10 and 20%, and under very high variability conditions even exceeding four times the purely statistical forecast. For the operationally relevant 15-min horizon, PV2-state exhibits a noteworthy precision, with two-thirds of forecasts falling within a 10% tolerance interval. Even under severe conditions roughly one third falls in this 10% interval.

Trade-offs in telemetry tag programming for deep-diving cetaceans: data longevity, resolution, and continuity

BackgroundAnimal-borne telemetry instruments (tags) have greatly advanced our understanding of species that are challenging to observe. Recently, non-recoverable instruments attached to cetaceans have increased in use, but these devices have limitations in data transmission bandwidth. We analyze trade-offs in the longevity, resolution, and continuity of data records from non-recoverable satellite-linked tags on deep-diving Ziphius cavirostris in the context of a behavioral response study of acute noise exposure. We present one data collection programming scheme that balances resolution and continuity against longevity to address specific questions about the behavioral responses of animals to noise exposure in experimental contexts. We compare outputs between two programming regimes on a commercially available satellite-linked tag: (1) dive behavior summary defined by conductivity thresholds and (2) depth time-series at various temporal resolutions.ResultsWe found that time-series data vary from the more precisely defined dives from a dive summary record data stream by an acceptable error range for our application. We determined a 5-min time-series data stream collected for 14 days balanced resolution with longevity, achieving complete or nearly complete diving records in 6 out of 8 deployments. We increased our data message reception rate several fold by employing a boat based data capture system. Finally, a tag deployed in a group concurrently with a high-resolution depth recorder showed high depth concordance.ConclusionsWe present the conceptual framework and iterative process for matching telemetry tag programming to research questions that we used and which should be applicable to a wide range of studies. Although designing new hardware for our specific questions was not feasible at the time, we were able to optimize the sampling regime of a commercially available instrument to meet the needs of our research questions and proposed analyses. Nevertheless, for other study species or designs, the complicated intersection between animal behavior and bandwidth of telemetry systems can often create a severe mismatch among research questions, data collection, and analysis tools. More flexible programming and purpose-built instruments will increase the efficacy of these studies and increase the scientific yield relative to the inherently higher risk of invasive studies.

Short-term PET-derived kinetic estimation for the diagnosis of hepatocellular carcinoma: a combination of the maximum-slope method and dual-input three-compartment model

BackgroundKinetic estimation provides fitted parameters related to blood flow perfusion and fluorine-18-fluorodeoxyglucose (18F-FDG) transport and intracellular metabolism to characterize hepatocellular carcinoma (HCC) but usually requires 60 min or more for dynamic PET, which is time-consuming and impractical in a busy clinical setting and has poor patient tolerance.MethodsThis study preliminarily evaluated the equivalence of liver kinetic estimation between short-term (5-min dynamic data supplemented with 1-min static data at 60 min postinjection) and fully 60-min dynamic protocols and whether short-term 18F-FDG PET-derived kinetic parameters using a three-compartment model can be used to discriminate HCC from the background liver tissue. Then, we proposed a combined model, a combination of the maximum-slope method and a three-compartment model, to improve kinetic estimation.ResultsThere is a strong correlation between the kinetic parameters K1 ~ k3, HPI and {{varvec{V}}}_{{varvec{b}}} in the short-term and fully dynamic protocols. With the three-compartment model, HCCs were found to have higher k2, HPI and k3 values than background liver tissues, while K1, k4 and {{varvec{V}}}_{{varvec{b}}} values were not significantly different between HCCs and background liver tissues. With the combined model, HCCs were found to have higher HPI, K1 and k2, k3 and {{varvec{V}}}_{{varvec{b}}} values than background liver tissues; however, the k4 value was not significantly different between HCCs and the background liver tissues.ConclusionsShort-term PET is closely equivalent to fully dynamic PET for liver kinetic estimation. Short-term PET-derived kinetic parameters can be used to distinguish HCC from background liver tissue, and the combined model improves the kinetic estimation.Clinical relevance statementShort-term PET could be used for hepatic kinetic parameter estimation. The combined model could improve the estimation of liver kinetic parameters.Graphical

Analyzing the impact of COVID-19 on the electricity demand in Austin, TX using an ensemble-model based counterfactual and 400,000 smart meters

The COVID-19 pandemic caused lifestyle changes and has led to the new electricity demand patterns in the presence of non-pharmaceutical interventions such as work-from-home policy and lockdown. Quantifying the effect on electricity demand is critical for future electricity market planning yet challenging in the context of limited smart metered buildings, which leads to limited understanding of the temporal and spatial variations in building energy use. This study uses a large scale private smart meter electricity demand data from the City of Austin, combined with publicly available environmental data, and develops an ensemble regression model for long term daily electricity demand prediction. Using 15-min resolution data from over 400,000 smart meters from 2018 to 2020 aggregated by building type and zip code, our proposed model precisely formalizes the counterfactual universe in the without COVID-19 scenario. The model is used to understand building electricity demand changes during the pandemic and to identify relationships between such changes and socioeconomic patterns. Results indicate the increase in residential usage , demonstrating the spatial redistribution of energy consumption during the work-from-home period. Our experiments demonstrate the effectiveness of our proposed framework by assessing multiple socioeconomic impacts with the comparison between the counterfactual universe and observations.

Nonlinear Regression-Based GNSS Multipath Dynamic Map Construction and Its Application in Deep Urban Areas

GNSS signals are easily blocked or degraded because of the dense presence of high-rise buildings in urban areas, and positioning errors arising from reflected signals amount to as much as hundreds of meters. Various conventional GNSS techniques have been utilized to resolve this problem, but applying them to urban environments has been difficult owing to the complexity of the reflected signals and their unpredictable and nonlinear variation to the signal receiving environments. In this study, multipath maps were generated for dynamic users at multiple positions on a road and a residual-based map selection algorithm was implemented to solve the problem of user position uncertainty in deep urban environments. GNSS data collected over a period of 327 min were used to train the multipath maps corresponding to 247 points near the 2.5 km stretch of the Teheran-ro road in Seoul, South Korea. The proposed system performed efficiently—it was verified to be capable of constructing a multipath map with a radius of 25 m using only 4 min of data. Moreover, it improved positioning accuracy by 45 % horizontally and by 80 % vertically, enabling the determination of the positional information of an urban vehicle with a horizontal accuracy of 18 m during 99% of the duration of a one-hour-long dynamic test. Due to the nonreliance of the proposed method on prior information or implementation of additional sensors, it is expected to be widely used for constructing map-based multipath mitigation models as part of intelligent transportation infrastructure in all cities in future.

Regional revenues of solar and wind generation in Texas

Effective regulatory policies, ample renewable energy potential, and a favorable business climate have led to a boom in wind farms and solar energy projects in Texas. Using a large sample of 15-min data for the six-year period of 2016–2021, we analyze per MWh revenues from solar generation and wind generation in the renewable regions of the Electric Reliability Council of Texas (ERCOT). We find that average regional revenues from variable renewable energy (VRE) for the daytime hours of 07:00–19:00 exceed the average levelized price of recently signed solar and wind power purchase agreements (PPAs), although average regional revenues for wind generation for the nighttime hours of 19:00–07:00 were lower than recent PPA prices. Moreover, VRE's 15-min regional revenues move with ERCOT's cap on the real-time market's energy price offers, regulatory price adder, daily wholesale natural gas price, 15-min nuclear energy generation, 15-min system load, 15-min regional solar energy and 15-min regional wind energy. While continued VRE development may suppress wholesale market prices, continued growth in electricity demand and increasing natural gas prices provide offsetting effects. Nevertheless, new policies are under development to mitigate the adverse effects of further VRE development on grid reliability and generation investment incentive.

Diagnosing the benign paroxysmal positional vertigo via 1D and deep-learning composite model.

Benign Paroxysmal Positional Vertigo (BPPV) is the leading cause of vertigo, and its characteristic nystagmus induced by positional maneuvers makes it a good model for Artificial Intelligence (AI) diagnosis. However, during the testing procedure, up to 10min of indivisible long-range temporal correlation data are produced, making the AI-informed real-time diagnosing unlikely in clinical practice. A combined 1D and Deep-Learning (DL) composite model was proposed. Two separate cohorts were recruited, with one for model generation and the other for evaluation of model's real-world generalizability. Eight features, including two head traces and three eye traces and their corresponding slow phase velocity (SPV) value, were served as the inputs. Three candidate models were tested, and a sensitivity study was conducted to determine the saliently important features. The study included 2671 patients in the training cohort and 703 in the test cohort. A hybrid DL model achieved a micro-area under the receiver operating curve (AUROC) of 0.982 (95% CI 0.965, 0.994) and macro-AUROC of 0.965 (95% CI 0.898, 0.999) for overall classification. The highest accuracy was observed for right posterior BPPV, with an AUROC of 0.991 (95% CI 0.972, 1.000), followed by left posterior BPPV, with an AUROC of 0.979 (95% CI 0.940, 0.998), the lowest AUROC was 0.928 (95% CI 0.878, 0.966) for lateral BPPV. The SPV was consistently identified as the most predictive feature in the models. If the model process is carried out 100 times for a 10-min data, one single running takes 0.79 ± 0.06s. This study designed DL models which can accurately detect and categorize the subtype of BPPV, enabling a quick and straightforward diagnosis of BPPV in clinical setting. The critical feature identified in the model helps expand our understanding of this disorder.

Clear-sky detection for PV degradation analysis using multiple regression

A method is presented to detect clear-sky periods for plane-of-array irradiance time-averaged data that is based on the algorithm originally described by Reno and Hansen. We show this new method improves the state-of-the-art by providing accurate detection at longer data averaging intervals. Moreover, our new method detects clear periods in plane-of-array data, which is novel. The new method is developed by applying a Design of Experiment approach to optimize the parameters used in the Reno method, and Monte Carlo simulations are used to understand the robustness of the found parameters. Clear-sky detection accuracy is compared among four methods: the Reno method, the default clear-sky filter in RdTools, the Ellis method, and the method outlined in this work, using a hand-labeled two-year data set of 1-min plane-of-array irradiance for a fixed tilt system. The RdTools clear-sky filter is marred by excessive false positives. The other methods all perform well at 1-min data intervals; the method developed here provides more accurate detection at longer data averaging intervals. We show that the parameters are directly linked to the data frequency in the hope that these input variables may not have to be optimized for every data frequency and location. However, only a single fixed system in one location was carefully examined. Finally, we illustrate how accurate determination of clear-sky conditions helps to eliminate data noise and bias in the assessment of long-term performance of PV plants.

Is Continuous Intraoperative Monitoring of Mean Arterial Pressure as Good as the Hypotension Prediction Index Algorithm?: Research Letter.

Is Continuous Intraoperative Monitoring of Mean Arterial Pressure as Good as the Hypotension Prediction Index Algorithm?: Research Letter.

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.