Time-varying Data Research Articles

Neural networks based multivariate time series forecasting of solar radiation using meteorological data of different cities of Bangladesh

Solar radiation is the energy or radiation we get from the sun, time-varying data. Solar radiation plays a vital role in various sectors. With better prediction, performances in these sectors can be enhanced. In this work, we proposed a system to forecast solar radiation using Neural Networks. Meteorological data from five different cities of Bangladesh were used. The system can forecast radiation values for any day using different meteorological data from the previous day. Three different networks were trained using the meteorological data, which are the Recurrent Neural Network (RNN), Long Short-Term Memory (LSTM), and Gated Recurrent Unit (GRU). Also, predictions were made for all five cities separately. An elaborate evaluation of all three models has been done to produce a comparison using widely used performance metrics. The GRU model produced the best result among all three models, with a MAPE score of 19.28%.

Solar Dish Stirling technology for sustainable power generation in Southern Morocco: 4-E analysis

This paper presents an energy, exergy, economic and environmental assessment of the Solar Dish Stirling (SDS) technology in thirteen Southern Moroccan sites, carrying out parametric analysis of various design parameters to optimise this technology under time-varying climatic data. System simulation was conducted on an hourly basis with the help of System Advisor Model (SAM) software. Annual energy can be maximised by adjusting the receiver aperture diameter, cooling pump speed, and thickness of receiver insulation appropriately. For all examined sites, the optimum value of the receiver aperture diameter and the thickness of receiver insulation were 70 mm and 30 mm, respectively. The recommended value of the pump speed varies depending on meteorological parameters for each site. The highest annual energy yield considering the optimal values of the assessed parameters is achieved in the site of Ouarzazate, which for a total installed capacity of 10 MW, reached a value of 23.693 GWh. Annual overall energetic and exergetic efficiencies were determined to be 24.26% and 25.98%, respectively. Nominal Levelized Cost of Electricity in the various cities varied in the range of 10.87c$/kWh and 18.69c$/kWh for Ouarzazate and Tan-Tan, respectively. Environmentally, the SDS technology can mitigate between 7.76 ktonne and 13.85 ktonne of CO2 emissions annually. Top three potential sites for implementing the Solar Dish Stirling technology were identified as Ouarzazate, Errachidia, and Midelt, assisting Moroccan policymakers in developing strategies to promote and plan future projects. Solar Dish Stirling technology can open the perspectives for modern polygeneration applications that can promote living conditions in Morocco’s southern regions.

Kimesurface Representation and Tensor Linear Modeling of Longitudinal Data.

Many modern techniques for analyzing time-varying longitudinal data rely on parametric models to interrogate the time-courses of univariate or multivariate processes. Typical analytic objectives include utilizing retrospective observations to model current trends, predict prospective trajectories, derive categorical traits, or characterize various relations. Among the many mathematical, statistical, and computational strategies for analyzing longitudinal data, tensor-based linear modeling offers a unique algebraic approach that encodes different characterizations of the observed measurements in terms of state indices. This paper introduces a new method of representing, modeling, and analyzing repeated-measurement longitudinal data using a generalization of event order from the positive reals to the complex plane. Using complex time (kime), we transform classical time-varying signals as 2D manifolds called kimesurfaces. This kime characterization extends the classical protocols for analyzing time-series data and offers unique opportunities to design novel inference, prediction, classification, and regression techniques based on the corresponding kimesurface manifolds. We define complex time and illustrate alternative time-series to kimesurface transformations. Using the Laplace transform and its inverse, we demonstrate the bijective mapping between time-series and kimesurfaces. A proposed general tensor regression based linear model is validated using functional Magnetic Resonance Imaging (fMRI) data. This kimesurface representation method can be used with a wide range of machine learning algorithms, artificial intelligence tools, analytical approaches, and inferential techniques to interrogate multivariate, complex-domain, and complex-range longitudinal processes.

Continual learning classification method with the weighted k-nearest neighbor rule for time-varying data space based on the artificial immune system

Supervised learning classification methods play an important role in various fields. However, most of them cannot effectively classify the data from time-varying data spaces, for they lack continual learning ability. Inspired by the intelligent mechanism that memory cells of the biological immune system can evolve with the evolution of invaders, a continual learning classification method with the weighted k-nearest neighbor rule for time-varying data space (WKNN-CLCMTVD) is proposed. Memory cells selected by the weighted k-nearest neighbor rule are used to identify the type of testing data. Memory cells are continuously cultivated, updated, and eliminated through learning testing data to improve the classification ability of WKNN-CLCMTVD. It degenerates into a standard supervised learning classification method when all data independent of time. Take experiments on twenty benchmark datasets, a 2-dimensional synthetic dataset, and XJTU-SY rolling element bearing accelerated life test datasets to evaluate the performance of WKNN-CLCMTVD. Results show that it has better classification ability for time-invariant data and outperforms the other methods for time-varying data space. Results also show that its running speed is far faster than that of other continual learning classification methods.

Assessment of Variations in Runoff Due to Landcover Changes Using the SWAT Model in an Urban River in Dublin, Ireland

Investigating the impact of land cover change in hydrological modelling is essential for water resources management. This paper investigates the importance of landcover change in the development of a physically-based hydrological model called SWAT. The study area considered is the Dodder River basin located in southern Dublin, Ireland. Runoff at the basin outlet was simulated using SWAT for 1993–2019 using five landcover maps obtained for 1990, 2000, 2006, 2012 and 2018. Results indicate that, in general, the SWAT model-simulated runoff for a chosen time-period are closer to the real-world observations when the landcover data used for simulation was collated as close to the time-period for which the simulations were performed. For 23 (20) years (from 27 years period) the monthly mean (maximum) runoff for the Dodder River generated by the SWAT model had the least error when the nearby landcover data were used. This study indicates the necessity of considering dynamic and time-varying landcover data during the development of hydrological modelling for runoff simulation. Furthermore, two composite quantile functions were generated by using a kappa distribution for monthly mean runoff and GEV distribution for monthly maximum runoff, based on model simulations obtained using different landcover data corresponding to different time-period. Modelling landcover change patterns and development of projected landcover in the future for river basins in Ireland needs to be integrated with SWAT to simulate future runoff.

Applying Deep Learning Model to Identify Heart Failure Patients with Kidney Failure Using Time-Varying Data

Applying Deep Learning Model to Identify Heart Failure Patients with Kidney Failure Using Time-Varying Data

Identification and Estimation of a Nonparametric Time-Varying Panel Data Model With Completely Missing Regressors in Some Periods

Identification and Estimation of a Nonparametric Time-Varying Panel Data Model With Completely Missing Regressors in Some Periods

A Novel Thermography-Based Dry Magnetic Particle Testing Method

Magnetic particle testing (MT) has been widely used in nondestructive testing (NDT) of ferromagnetic materials. MT can be divided into wet method and dry method. The wet method is highly sensitive for microcracks and easy to operate. But the problem is that magnetic suspension liquid utilized in the wet method pollutes both test objects and environment. Dry magnetic particles utilized in dry method are recyclable and environmentally friendly. But the irrelevant magnet particles must be manually removed to reveal the magnetic particle indication (MPI). This operation limits both the detection efficiency and sensitivity of dry method. To increase the automation in MT, this article proposes a thermography-based dry MT method (TDMT). Specifically, the dry magnetic particles are heated before utilization and then thermal data are utilized for crack identification. TDMT inherits the economic and environmental advantages of the dry method. Additionally, the time-varying thermal data can provide more useful information than conventional MPI, which makes it possible to identify cracks without manual assistance. The mathematical model of TDMT is designed to prove its feasibility and a dimensionality reduction-based algorithm is applied for crack signal enhancement. Experiments on different specimens with artificial and natural cracks have been conducted. The notch with both a depth of <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$15~\mu \text{m}$ </tex-math></inline-formula> on the standard test shim and a width up to 0.8 mm on the steel plate can be successfully detected. The high sensitivity and wide detection range can be seen from TDMT.

RoboHapalytics: A Robot Assisted Haptic Controller for Immersive Analytics.

Immersive environments offer new possibilities for exploring three-dimensional volumetric or abstract data. However, typical mid-air interaction offers little guidance to the user in interacting with the resulting visuals. Previous work has explored the use of haptic controls to give users tangible affordances for interacting with the data, but these controls have either: been limited in their range and resolution; were spatially fixed; or required users to manually align them with the data space. We explore the use of a robot arm with hand tracking to align tangible controls under the user's fingers as they reach out to interact with data affordances. We begin with a study evaluating the effectiveness of a robot-extended slider control compared to a large fixed physical slider and a purely virtual mid-air slider. We find that the robot slider has similar accuracy to the physical slider but is significantly more accurate than mid-air interaction. Further, the robot slider can be arbitrarily reoriented, opening up many new possibilities for tangible haptic interaction with immersive visualisations. We demonstrate these possibilities through three use-cases: selection in a time-series chart; interactive slicing of CT scans; and finally exploration of a scatter plot depicting time-varying socio-economic data.

A Synchronous Detection Algorithm for Quasi-Periodic Signal Components and Its Application in Photoplethysmographic Imaging

photoplethysmographic imaging (PPI) is a noninvasive method for assessing local blood flow beneath the skin surface by a video camera. The synchronization detection algorithm based on lock-in amplification (LIA) is a widely used method to obtain PPI. This method can obtain two maps: one is the amplitude map, which represents the strength of the PPG signal, and the other is the pulse-transit-time (PTT) map, which represents the time synchronization of the PPG signal. The latter is usually represented directly in form of a phase map. However, in general, the frequency of PPG signals is time-varying and contains harmonic components, which complicates the application of the algorithm. This article analyzes the relationship between the LIA calculation results and the time delay when the frequency of the signal is time-varying or the signal contains harmonics. Theoretical analysis results show that, when the signal contains harmonic components, the relationship between time delay and phase is approximately linear only when the time delay is not very large or the harmonic components are not very strong. Meanwhile, amplitude appears as a “bell-shaped” function, with its peak value located at the zero time delay. Video data were acquired by the video PPG acquisition system built in our laboratory, and the PPI was calculated by the synchronous detection algorithm using a reference signal generated directly by the Hilbert transform of an actual signal. Experiments show that the proposed method can obtain more stable results than using a constant-frequency reference signal when dealing with time-varying heart rate data. An interesting finding is that, when the reference signal contains harmonic components, PPG distribution features in the amplitude map can be enhanced.

Dynamic Bayesian Network Based Security Analysis for Physical Layer Key Extraction

Internet of Things (IoT) is envisioned to expand Internet connectivity of the physical world, and the mobile edge cloud can be leveraged to enhance the resource-constrained IoT devices. The performance of the cloud-enhanced IoT applications depends on various system-wide information, such as the wireless channel states between IoT devices and their corresponding serving edge cloud nodes. However, with the semi-trusted edge resources and the public nature of wireless channels, public sharing of system information should be avoided to better balance the tradeoff between performance and security. In this paper, the benefits of local information exchange is investigated, where the privately-owned physical layer channel information is leveraged to extract lightweight keys. For the point-to-point wireless communications links with multiple passive eavesdroppers, the security metric in terms of conditional min-entropy is evaluated via the proposed Dynamic Bayesian Model. The proposed model can flexibly incorporate various dynamic information flows in the system and quantify the information leakage caused by wireless broadcasting. The rigorously defined and derived security metrics for such a key generation pipeline has been verified via the real-world collected time-varying wireless channel data. The designed model can achieve previously inconceivable security properties.

A feature visualization method for time-varying volume data

A feature visualization method for time-varying volume data

Sparse and Time-Varying Predictive Relation Extraction for Root Cause Quantification of Nonstationary Process Faults

Root cause diagnosis (RCD) is an important technique for maintaining process safety, which infers the causalities between faulty measurements to locate the root cause of the fault. However, process nonstationarity brings time-varying data properties, preventing RCD models from accurately revealing causality. Although existing models can capture time-varying predictive relations among nonstationary variables to characterize causalities, they cannot apply nonsmooth sparsity constraints to the variable selection during parameter updating, leading to redundancy relations. To overcome this challenge, a novel sparse and time-varying predictive relation extraction method is proposed for the RCD of nonstationary processes. First, a time-varying causal matrix is constructed to guide information fusion among variables in the prediction task under sparsity constraints, thus capturing significant causal dependencies. Second, a novel parameter updating algorithm enhanced by the Taylor expansion and subgradient techniques is designed for the causal matrix. Theoretical verification shows that the designed algorithm can obtain sparse and closed-form solutions, overcoming the non-smoothness challenge and ensuring the causal significance. Finally, a causal stability metric is developed to screen reliable causalities from time-varying relations, further eliminating redundant causalities by considering both average strength and variability information. The validity of the proposed method is illustrated through both the Tennessee Eastman benchmark example and a real industrial process.

A Hybrid Spiking Neurons Embedded LSTM Network for Multivariate Time Series Learning under Concept-drift Environment

Complicated temporal patterns can provide important information for accurate time series forecasting. Existing long short-term memory (LSTM) model with attention mechanism have achieved significant performance. However, the exponential decay of long-term memory of LSTM has not be resolved yet in these efforts, remaining a longstanding open problem in recurrent nature. This problem exhibits a bottleneck which restricts the performance of existing studies. Recently, spiking neural networks (SNNs) have shown high efficiency in capturing temporal patterns via the surrogate gradient (SG) method to resolve this issue. However, the concept-drift environment makes it impossible to pre-set the variance into the standard SG method due to time-varying data distribution. In this paper, we propose a novel adaptive and hybrid spiking (AHS) module embedded LSTM, collaborating with two attention mechanisms (called HSN-LSTM) to resolve above-mentioned problems. First, the AHS module is analyzed theoretically can remain long-term memory. Moreover, our smooth SG method avoids pre-setting of variance, which is not sensitive in the above scenarios. Besides, we use the negative log-likelihood function to adjust the attention score for alleviating the negative impact from the concept-drift. Experiment results show the HSN-LSTM outperformed the state-of-the-art models on several multivariate time series datasets.

STNet: An End-to-End Generative Framework for Synthesizing Spatiotemporal Super-Resolution Volumes.

We present STNet, an end-to-end generative framework that synthesizes spatiotemporal super-resolution volumes with high fidelity for time-varying data. STNet includes two modules: a generator and a spatiotemporal discriminator. The input to the generator is two low-resolution volumes at both ends, and the output is the intermediate and the two-ending spatiotemporal super-resolution volumes. The spatiotemporal discriminator, leveraging convolutional long short-term memory, accepts a spatiotemporal super-resolution sequence as input and predicts a conditional score for each volume based on its spatial (the volume itself) and temporal (the previous volumes) information. We propose an unsupervised pre-training stage using cycle loss to improve the generalization of STNet. Once trained, STNet can generate spatiotemporal super-resolution volumes from low-resolution ones, offering scientists an option to save data storage (i.e., sparsely sampling the simulation output in both spatial and temporal dimensions). We compare STNet with the baseline bicubic+linear interpolation, two deep learning solutions ( SSR+TSF, STD), and a state-of-the-art tensor compression solution (TTHRESH) to show the effectiveness of STNet.

Capturing dynamics of time-varying data via topology

&lt;p style='text-indent:20px;'&gt;One approach to understanding complex data is to study its shape through the lens of algebraic topology. While the early development of topological data analysis focused primarily on static data, in recent years, theoretical and applied studies have turned to data that varies in time. A time-varying collection of metric spaces as formed, for example, by a moving school of fish or flock of birds, can contain a vast amount of information. There is often a need to simplify or summarize the dynamic behavior. We provide an introduction to topological summaries of time-varying metric spaces including vineyards [&lt;xref ref-type="bibr" rid="b19"&gt;19&lt;/xref&gt;], crocker plots [&lt;xref ref-type="bibr" rid="b55"&gt;55&lt;/xref&gt;], and multiparameter rank functions [&lt;xref ref-type="bibr" rid="b37"&gt;37&lt;/xref&gt;]. We then introduce a new tool to summarize time-varying metric spaces: a &lt;i&gt;crocker stack&lt;/i&gt;. Crocker stacks are convenient for visualization, amenable to machine learning, and satisfy a desirable continuity property which we prove. We demonstrate the utility of crocker stacks for a parameter identification task involving an influential model of biological aggregations [&lt;xref ref-type="bibr" rid="b57"&gt;57&lt;/xref&gt;]. Altogether, we aim to bring the broader applied mathematics community up-to-date on topological summaries of time-varying metric spaces.&lt;/p&gt;

By What Extent the Japanese Has Been Surprised by COVID-19 Information? Emotional Contagion and its Mathematical Model

A strong field of negative information as the case realized by COVID-19 must greatly enhance the public feelings such as the fear, the disgust and a sense of crisis against it. The information environment regarding COVID-19 and the public reactions to it, which had appeared from the early 2020 to September 2021 in Japan, were first reviewed. The proposition of a mathematical model was followed, where the negative and strong field of information as the issue of COVID-19 was assumed to make the public surprised and their emotions changed negatively via the mechanism of emotional contagion. The frequency of the public access to SNS and Internet under the atmosphere of COVID-19 was considered as a manifestation of the public for seeking reliable information and as an index of the public reaction originated from the negative emotion. By using the time-varying data of the information field of COVID-19, the extent of the public surprise that was considered as the emotional contagion from the field was derived. That extent became clear to have varied heterogeneously with time, having been subjected by the state of information field which have varied with three phases as the initial, transient and quasi-stable ones. It was found that, only in the quasi-stable state of the field, the number of public access to SNS and Internet can mimic the social reality of the issue concerned. Physical and mathematical models, where the psychological phenomena as the emotional contagion are positively included, were pointed out to be central ingredients to understand the behavior of the public in the society more comprehensively than it is today.

A Comparison of Spatiotemporal Visualizations for 3D Urban Analytics.

Recent technological innovations have led to an increase in the availability of 3D urban data, such as shadow, noise, solar potential, and earthquake simulations. These spatiotemporal datasets create opportunities for new visualizations to engage experts from different domains to study the dynamic behavior of urban spaces in this under explored dimension. However, designing 3D spatiotemporal urban visualizations is challenging, as it requires visual strategies to support analysis of time-varying data referent to the city geometry. Although different visual strategies have been used in 3D urban visual analytics, the question of how effective these visual designs are at supporting spatiotemporal analysis on building surfaces remains open. To investigate this, in this paper we first contribute a series of analytical tasks elicited after interviews with practitioners from three urban domains. We also contribute a quantitative user study comparing the effectiveness of four representative visual designs used to visualize 3D spatiotemporal urban data: spatial juxtaposition, temporal juxtaposition, linked view, and embedded view. Participants performed a series of tasks that required them to identify extreme values on building surfaces over time. Tasks varied in granularity for both space and time dimensions. Our results demonstrate that participants were more accurate using plot-based visualizations (linked view, embedded view) but faster using color-coded visualizations (spatial juxtaposition, temporal juxtaposition). Our results also show that, with increasing task complexity, plot-based visualizations perform better in preserving efficiency (time, accuracy) compared to color-coded visualizations. Based on our findings, we present a set of takeaways with design recommendations for 3D spatiotemporal urban visualizations for researchers and practitioners. Lastly, we report on a series of interviews with four practitioners, and their feedback and suggestions for further work on the visualizations to support 3D spatiotemporal urban data analysis.

Longitudinal effects of physical and social neighbourhood change on suicide mortality: A full population cohort study among movers and non-movers in the Netherlands

Associations between the residential neighbourhood environment and suicide mortality are well-established; however, most evidence is cross-sectional and not capable of incorporating place-based and residential moving-related neighbourhood changes. We studied how suicide mortality is associated with changes in the physical and social neighbourhood environment for movers and non-movers. Our retrospective analysis was based on longitudinal register data for the entire Dutch population aged 25–64 years enriched with annually time-varying data on the residential neighbourhood environment between 2007 and 2016. A total of 8,741,021 people were followed-up between 2007 and 2016 of which 10,019 committed suicide. Upward and downward neighbourhood change was measured by comparing neighbourhood conditions separately at two time points. Cox proportional hazard models indicated that movers had a significantly lower risk of suicide compared to non-movers. Suicide risk was lower for people experiencing improvements in social fragmentation and deprivation compared to those remaining in poor conditions. Change from rural to urban conditions also resulted in lower suicide risk, while a gain in green space put people at increased risk. For those stable neighbourhood conditions over time, suicide mortality was lower for men and women in urban vs. rural neighbourhoods as well as for women in neighbourhoods with low vs. high social fragmentation. Stable exposure to high levels of green space resulted in higher suicide risk among women. Interactions and stratification by moving type revealed associations between neighbourhood change and suicide were more pronounced in non-movers. Our findings suggest that neighbourhood improvements might contribute to a lower suicide risk, especially for long-term residents in poor neighbourhood conditions.

Quantification of ecosystem services supply-demand and the impact of demographic change on cultural services in Shenzhen, China

Previous studies on the supply and demand of ecosystem services (ES) mainly focused on inter-annual changes, and no studies have explored the impact of demographic change on the ES supply and demand on fine-grained time scales. Thus, taking Shenzhen as an example, the status of ES supply and demand, as well as diurnal population changes and their impacts on cultural services were analyzed at different time periods using mobile phone signaling data, ecological supply-demand ratio (ESDR), Geo-Informatic Tupu, InVEST model and buffer zone. The results showed that the population declines successively on workdays, weekends and holidays, and that the daytime population is greater than the nighttime. Water yield services can basically meet the demand in terms of quantity and spatial distribution, however, carbon sequestration and cultural services showed the opposite results. The main type of ESDR changes in cultural services are the mutual conversion of deficit and balance, and these are concentrated in areas with high forest coverage and small populations, but frequent population changes. In addition, when the fixed population is too large, the use of time-varying population data will conceal the impact of demographic changes on ES supply and demand, and other data are needed for auxiliary analysis. Overall, this study provides a new research perspective for the ES supply and demand and can provide a theoretical basis for refined sustainable urban management.