Dynamic Time Research Articles

Differences in Changes of Mobility Patterns Across the Globe—Evidence From a Natural Experiment

ABSTRACTThe impact of population mobility and transportation choices on the environment is significant. Sustainable mobility policies require an understanding of evolving mobility patterns. This study examines global mobility variations across countries using COVID‐19 Community Mobility Reports. Employing the k‐medoids algorithm and Dynamic Time Warping, we analyzed mobility dynamics. Results reveal diverse changes in population mobility. Around 52 Global North countries exhibit approximately 10% reduction in professional activity‐related traffic post‐COVID‐19. Regarding urban green space mobility, only 29 countries exhibit strong seasonality, with summer traffic in the northern hemisphere peaking at about 150% higher than winter traffic. Three groups of countries are identified concerning public transport mobility: returning to pre‐pandemic levels, experiencing a 25% increase, and nearly doubling from pre‐pandemic levels. This underscores the key determinants for sustainable mobility policy implementation. Fifteen highly developed countries share similar mobility patterns across six areas studied, facilitating the exchange of sustainable mobility solutions and best practices. This research underscores the importance of understanding and addressing evolving mobility patterns for effective environmental policy planning.

Performing Dynamic Analysis of the Conceret Structures using Genetic Algorithm and Determining the Optimal Range during a Natural Hazard Crisis (Tempe, Arizona, USA)

The approach employs a binary genetic algorithm alongside natural numbers. Given the differences between accelerometers and scale coefficients, the genetic algorithm introduced herein is a hybrid, featuring two chromosomes. A key factor influencing the accuracy and effectiveness of these programs is the precise estimation of their parameters. When these parameters are determined accurately, the gap between the average response spectrum and the design spectrum is significantly minimized. This program operates with two genetics that run concurrently, yielding results that closely approximate the optimal solution. The program itself is capable of offering users a range of desired coefficients, as well as values for the covariance and mutation of each chromosome. The algorithm detailed in this article can generate a new generation of individuals from an infinite array of ground motion records, utilizing processes such as natural selection, mating, and mutation, and continue this cycle until an individual meets the desired characteristics. This paper presents a novel method for optimally selecting accelerometers and scaling them for dynamic time history analysis, aiming for an average response spectrum that closely aligns with the target spectrum, reflecting the expected seismic activity for the structure in Tempe district in Arizona State. Given the relatively high number of parameters involved, reliance on trial-and-error methods is significantly influenced by the user’s skill set; thus, the hybrid genetic algorithm presented here addresses this limitation.

Multivariate time-series clustering of cardiopulmonary exercise test data for cardiovascular risk stratification

Abstract Aims Current clinical practice assesses cardiorespiratory fitness (CRF) by evaluating summary indexes of cardiopulmonary exercise tests (CPET). Yet, the raw time series recordings may hold additional information with clinical relevance. In this study, we investigated if the interpretation of raw CPET data could identify distinct CRF phenogroups and improve risk stratification. Methods In 1399 participants (mean age, 56.4 years; 37.7% women), who underwent maximal CPET by cycle ergometer, we collected baseline clinical characteristics and information on incident cardiovascular (CV) events (n=297) on average 4.3 years later. We employed dynamic time warping combined with k-medoids to identify phenogroups from raw CPET time-series. To evaluate the clinical significance of the derived phenogroups we compared clinical characteristics and incidence of CV events, while an external population cohort (n=171) was used to further validate the trained model. Results The optimal number of clusters was 5. We observed significant differences in age, use of medication, spirometry indexes and disease prevalence across all clusters. Cluster 5 was associated with the worst CV profile with higher use of antihypertensive medication and history of CV disease while cluster 1 represented the most favourable risk profile. After adjusting for traditional clinical covariables, clusters 4 (HR: 1.30; 95%CI: 1.02-1.65; p=0.033) and 5 (HR: 1.36; 95%CI: 1.08-1.72; p=0.0088) had significantly higher risk for incident CV events compared to clusters 1 and 2. Comparing clinical characteristics across clusters in the external validation cohort revealed similar patterns. Conclusion – Employing unsupervised machine learning on time series CPET recordings, we identified clinically meaningful CRF phenogroups. An integrative CRF profiling might facilitate CV risk stratification and consequently risk management.

Early Internal Short Circuit Diagnosis for Lithium-Ion Battery Packs Based on Dynamic Time Warping of Incremental Capacity

Timely identification of early internal short circuit faults, commonly referred to as micro short circuits (MSCs), is essential yet poses significant challenges for the safe and reliable operation of lithium-ion battery (LIB) energy storage systems. This paper introduces an innovative diagnostic method for early internal short circuits in LIB packs, utilizing dynamic time warping (DTW) applied to incremental capacity (IC). Initially, the terminal voltages of all cells within the LIB pack are ordered at any moment to determine the median terminal voltage, which is then used to generate the median IC curve. This curve acts as a reference benchmark that represents the condition of healthy cells in the pack. Subsequently, the DTW algorithm is utilized to measure the similarity between each cell’s IC curve and the median IC curve. Cells exhibiting similarity scores that exceed a specified threshold are identified as having MSC faults. Lastly, for the cells diagnosed with MSC conditions, a method for estimating short-circuit resistance (SR) based on variations in maximum charging voltage is devised to quantitatively evaluate the severity and evolution of the MSC. Experimental findings reveal that the proposed method effectively identifies MSC cells in the LIB pack and estimates their SRs without the necessity of a battery model, thereby affirming the method’s validity.

Time-series clustering of raw recordings of 24-hour ambulatory blood presssure for risk stratification in general population

Abstract Objective 24-hour ambulatory blood pressure measurements (ABPM) is the best out-of-office BP measurement used to refine cardiovascular (CV) risk stratification. Currently, only summary ABPM indexes are used in the clinical practice while the temporal characteristics of the raw BP recordings have not been explored. Therefore, in this study we employed an unsupervised machine learning (ML) algorithm suitable for multivariate time-series analysis to identify distinct BP and heart rate (HR) profiles associated with the incidence of adverse CV events. Design and Method In 1391 community-dwelling individuals (mean age, 46.9 years; 50.97% women), we acquired 24-hour ABPM and clinical data and collected CV events (n=399) on average 22 years later. We employed multivariate dynamic time warping and k-medoids algorithm on the raw recordings of systolic and diastolic BP and HR. To assess the clinical significance of the derived clusters, we compared the clinical characteristics and the incidence of adverse events. We validated the trained model using 24-hour ABPM obtained from external population cohort (n=1137). Results The silhouette score and Dunn index showed that the optimal number of clusters was 4. Across all clusters we observed significant difference in age and sex distribution. Cluster 4 had the worst CV risk factors profiling, with higher office BP and anti-hypertensive medication intake compared to the other clusters. We observed differences in the incidence of adverse events between clusters (Figure), with cluster 1 representing the lowest risk group and cluster 4 the highest. After adjusting for traditional CV risk factors, including office systolic BP, cluster 1 had the lowest risk (HR:0.92; 95%CI:0.84-0.99; P=0.036) and cluster 4 the highest (HR:1.13; 95%CI:1.03-1.24; P=0.006) as compared to the average population risk. The analysis of the external validation cohort revealed similar ABPM patterns and clinical characteristics. Conclusion Employing an unsupervised ML model on the raw ABPM recordings we identified clinically meaningful clusters associated with an increasing cumulative incidence of CV events. This ML analysis paves the way towards utilization of temporal BP and HR information and subsequently optimization of CV risk stratification.

Structural Plasticity as a Driver of the Maturation of Pro-Interleukin-18.

Dynamics are often critical for biomolecular function. Herein we explore the role of motion in driving the maturation process of pro-IL-18, a potent pro-inflammatory cytokine that is cleaved by caspases-1 and -4 to generate the mature form of the protein. An NMR dynamics study of pro-IL-18, probing time scales over 12 orders of magnitude and focusing on 1H, 13C, and 15N spin probes along the protein backbone and amino-acid side chains, reveals a plastic structure, with millisecond time scale dynamics occurring in a pair of β-strands, β1 and β*, that show large structural variations in a comparison of caspase-free and bound pro-IL-18 states. Fits of the relaxation data to a three-site model of exchange showed that the ground state secondary structure is maintained in the excited conformers, with the side chain of I48 that undergoes a buried-to-exposed conformational change in the caspase-free to -bound transition of pro-IL-18, sampling a more extensive range of torsion angles in one of the excited states characterized, suggesting partial unpacking in this region. Hydrogen exchange measurements establish the occurrence of an additional process, whereby strands β1 and β* locally unfold. Our data are consistent with a hierarchy of dynamic events that likely prime pro-IL-18 for facile caspase binding.

A simple dynamic time constant model for both summer overheating and winter heating analysis

As new buildings have high levels of insulation it is important to ensure that solar and other internal gains do not make the interiors too hot in summer. This is well known. But assessing overheating at an early design stage is not so easy. The Thermal Mass Parameter (TMP) in the regulations is based on a 24 h cyclic assessment of the depth of heat storage in a material. This paper examines a simpler method using the whole thermal mass, simple steady state calculations and a time constant that can take into account the heat gains over a number of days, such as in a heat wave, not just over 24 h. This simple dynamic time constant model (SDTCM) would be a more realistic model to assess overheating and performance for Future Homes and Buildings Standards, now under consultation in the UK. The model can equally be applied to assess winter heating and for dynamic simulation to consider control strategies. Practical application This paper describes a simple model that can be used for determining heating sizing in winter and overheating in summer. It relates to assessing the thermal mass via the time constant relating to the dynamic response of the building. This will be useful in initial design and possibly regulation.

Modeling Ordinal First Return Dynamics from Time Series

We examine the family of ordinal First Return Maps (FRMs) of a dynamical system time series to model the system’s dynamics. We consider two distinct modeling approaches using the same modeling method and compare their outcomes to ascertain the prerequisites for each to function and which approach is more effective in different circumstances. The first approach involves short-term modeling of the entire flow, followed by identifying the ordinal sections on the predicted flow. The second approach relies on locating the ordinal sections within the reference time series and using the data from each section independently to build a model that can produce a long-term prediction for each section’s FRM. We demonstrate the effectiveness of these approaches on two well-known dynamical systems, the Lorenz and Rössler equations, emulating data deficiency by selecting a low sampling rate. We consider the effect of downsampling on the shape of the FRMs to choose suitable ordinal sequences to use in creating Poincaré sections from the time series. Following that, we propose a dissimilarity measure between the reference and predicted FRMs to assess the model quality. We then compare the reference and predicted FRMs using the introduced dissimilarity criterion to determine which modeling approach performs better in each situation. Ultimately, we classify the FRMs situation into continuous and piecewise continuous scenarios. Based on our observations from the Lorenz and Rössler dynamical systems, we found that when the system’s FRMs are piecewise continuous, it is more effective to model the entire flow. However, when the FRMs are continuous, our choice of modeling approach depends on the criterion that the predicted FRM is more similar to the reference.

CHANG-ES. XXXIII. A 20 kpc radio bubble in the halo of the star-forming galaxy NGC 4217

Cosmic rays may be dynamically very important in driving large-scale galactic winds. Edge-on galaxies give us an outsider's view of radio haloes, and of their extra-planar cosmic-ray electrons and magnetic fields. We present a new radio continuum imaging study of the nearby edge-on galaxy NGC\,4217. We examine the distribution of extra-planar cosmic rays and magnetic fields. We observed it with both the Jansky Very Large Array (JVLA) in the $S$ band (2--4\,GHz) and the LOw Frequency ARray (LOFAR) at 144\,MHz. We measured vertical intensity profiles and exponential scale heights. We re-imaged both the JVLA and LOFAR data at matched angular resolution in order to measure radio spectral indices between 144\,MHz and 3\,GHz. Confusing point-like sources were subtracted prior to imaging. We then fitted intensity profiles with cosmic-ray electron advection models, using an isothermal wind model that is driven by a combination of pressure from the hot gas and cosmic rays. We discover a large-scale radio halo on the north-western side of the galactic disc. The morphology is reminiscent of a bubble extending up to 20\,kpc from the disc. We find spectral ageing in the bubble, which allowed us to measure the advection speeds of the cosmic-ray electrons, which accelerate from 300 to $600\ $. Assuming energy equipartition between the cosmic rays and the magnetic field, we estimate the bubble may have been inflated by a modest 10\,<!PCT!> of the kinetic energy injected by supernovae over its dynamical timescale of 35\,Myr. While no active galactic nucleus (AGN) has been detected, such activity in the recent past cannot be ruled out. Non-thermal bubbles with sizes of tens of kiloparsecs may be a ubiquitous feature of star-forming galaxies, and if so this would demonstrate the influence of feedback. Determining possible contributions by AGN feedback will require deeper observations.

Acoustic Scene Classification using Dynamic Time Warping Technique based on Short Time Fourier Transform and Discrete Wavelet Transforms

Acoustic Scene Classification using Dynamic Time Warping Technique based on Short Time Fourier Transform and Discrete Wavelet Transforms

A dynamically consistent discretization method for the Goodwin model with nonlinear Phillips curve. Comparing qualitative and quantitative dynamics

AbstractThe Goodwin model is a widely used economic growth model able to explain endogenous fluctuations in employment rate and wage share; in its initial version, the standard Phillips curve is used. In the present work, we suggest a revised Phillips curve that takes into account how the wage share influences the rate of changes of the wage itself thus obtaining a continuous-time modified Goodwin model. Since applying models to real data often requires working in a discrete-time setup, we then move from the continuous-time to the discrete-time version of the proposed model, by using a general polynomial discretization method in backward and forward-looking (hybrid discretization). By comparing the continuous-time system to its discrete-time counterpart we prove that fixed points and local dynamics do not change, as long as the time step is not too high. Moreover, numerical simulations employing Dynamic Time Warping, cross-correlation, and semblance analysis consistently affirm that enhancing the similarity of quantitative dynamics is achieved by reducing the time step.

Enhancing underwater target detection: Fusion of spatio‐temporal incompletely‐aligned AIS and sonar information via DTW and multi‐head attention mechanism

AbstractIn the field of underwater target detection, the passive sonar is an important means of long‐distance target detection. The sonar detection information typically includes both surface and underwater targets, whereas it is a great challenge on effectively distinguishing between surface and underwater targets solely based on sonar information. Effective fusion of sonar and AIS (Automatic Identification System) data can leverage their complementary nature to compensate for the limitation of sonar information. However, the sonar information and AIS information are acquired based on different detection principles and systems, which are essentially multi‐source heterogeneous information with obvious spatio‐temporal misalignment in nature. Existing fusion methods normally struggle to effectively align sonar and AIS data in both time and space subject to the complexity of the problem. In this study, the Dynamic Time Warping (DTW) algorithm is applied to align sonar and AIS data in the time domain. In addition, a deep learning algorithm with multi‐head attention mechanism is proposed to achieve the spatial alignment of sonar and AIS data, where the matching between the surface targets in AIS data and the same surface targets in sonar data can also be successfully achieved. It provides a priori knowledge to enhance the underwater target detection of the passive sonar by eliminating the interference of the surface targets. Based on the attention mechanism, the abstract features extracted from the intermediate‐layer of the neural networks are found to be effective to represent the typical features of the target motion trajectories, which also demonstrates the effectiveness of the attention mechanism. The experiment results show that the proposed method can successfully achieve a MatchingSucccessRate of over 95% between the AIS targets and sonar detection targets.

A Cluster-Based Deep Learning Model Perceiving Series Correlation for Accurate Prediction of Phonon Spectrum.

The spectral properties are the most prevalent continuous representation for characterizing transport phenomena and excitation responses, yet their accurate predictions remain a challenge due to the inability to perceive series correlations by existing machine learning (ML) models. Herein, a ML model named cluster-based series graph networks (CSGN) is developed based on the dynamical theory of crystal lattices to predict phonon density of states (PDOS) spectrum for crystal materials. The multiple atomic cluster representation is constructed to capture the diverse vibration modes, while the mixture Gaussian process and dynamic time warping mechanism are compiled to project from clusters to PDOS spectrum. Accurate predictions of complicated spectra with multiple or overlapping peaks are achieved. The high performance of CSGN model can be attributed to the pertinent feature extraction and the appropriate similarity evaluation, which enable the natural perception of structure-property relation and intrinsic series correlations as confirmed in the predictive results. The transferable and interpretable CSGN model advances ML predictions of spectral properties and reveals the potential of designing ML methods based on physical mechanisms.

Effects of inhaled beclometasone dipropionate/formoterol fumarate/glycopyrronium vs. beclometasone dipropionate/formoterol fumarate and placebo on lung hyperinflation and exercise endurance in chronic obstructive pulmonary disease: a randomised controlled trial

BackgroundThe single-inhaler triple combination of beclometasone dipropionate, formoterol fumarate, and glycopyrronium (BDP/FF/G) is available for maintenance therapy of chronic obstructive pulmonary disease (COPD). Cardinal features of COPD are lung hyperinflation and reduced exercise capacity. TRIFORCE aimed to evaluate the effect of BDP/FF/G on lung hyperinflation and exercise capacity in patients with COPD.MethodsThis double-blind, randomised, active- and placebo-controlled, crossover study recruited adults with COPD aged ≥ 40 years, who were hyperinflated and symptomatic, and were receiving mono- or dual inhaled maintenance COPD therapy. In the three treatment periods, patients were randomised to receive BDP/FF/G, BDP/FF, or placebo, each for 3 weeks, with a 7–10-day washout between treatment periods. Assessments included slow inspiratory spirometry (for resting inspiratory capacity [IC]) and constant work-rate cycle ergometry (for dynamic IC and exercise endurance time). The primary objective was to compare BDP/FF/G and BDP/FF vs. placebo for resting IC at Week 3. Key secondary objectives were to compare BDP/FF/G and BDP/FF vs. placebo for dynamic IC and exercise endurance time during constant work rate cycle ergometry at Week 3.ResultsOf 106 patients randomised, 95 completed the study. Resting IC adjusted mean differences vs. placebo were 315 and 223 mL for BDP/FF/G and BDP/FF, respectively (p < 0.001 for both). Adjusted mean differences vs. placebo for the key secondary endpoints were: 245 mL for dynamic IC (p < 0.001) and 69.2 s for exercise endurance time (nominal p < 0.001) with BDP/FF/G, and 96 mL (p = 0.053) and 70.1 s (nominal p < 0.001) with BDP/FF. Differences between BDP/FF/G and BDP/FF for resting and dynamic IC were 92 and 149 mL (p < 0.01 for both). All three treatments were generally well tolerated, with 27.3%, 25.3% and 19.0% of patients reporting adverse events with BDP/FF/G, BDP/FF and placebo, respectively, all mild or moderate.ConclusionsIn patients with COPD, BDP/FF/G provided significant and clinically relevant improvements vs. placebo and BDP/FF in static and dynamic hyperinflation, with an improvement vs. placebo in exercise endurance.Trial registrationClinicalTrials.gov (NCT05097014), registered 27th October 2021.

A robust method for mapping soybean by phenological aligning of Sentinel-2 time series

Soybean is an important crop for food and animal feed. Production and area both continue to increase and expand into new areas and countries. Spatially explicit information on soybean cultivation is essential to crop monitoring, production estimation, and national accounting systems. However, its cultivation in diverse climate conditions, landscapes, and agricultural systems poses challenges to accurately map soybean across different regions and years. We propose an innovative soybean mapping method combining phenological alignment with machine learning (named here RF-DTW), which can be applied to diverse geographies and years by aligning phenological shifts and using distinctive features from Sentinel-2 time-series. The method first uses the dynamic time warping (DTW) algorithm to align the growing season between pixels across different sites. Then, based on the harmonized time-series, a set of distinctive features was identified and used to build random forest (RF) models to classify soybean across ten globally distributed sites and multiple years. Results show that the green chlorophyll vegetation index (GCVI), greenness and water content composite index (GWCCI), normalized difference senescent vegetation index (NDSVI), red edge position (REP), and short-wave infrared bands are important inputs for distinguishing soybean from other crops. Spectral-phenological features, particularly the curve slope metrics of GCVI and GWCCI during the peak to late growing season, rank as the most important features for mapping soybean. RF-DTW demonstrates good generalizability across ten study sites, achieving an overall accuracy (OA) of 0.92 and an F1-score of 0.84. F1-scores for eight out of ten sites ranged between 0.82 and 0.98, outperforming a benchmark method, although they were lower (F1-score < 0.60) for the two sites in Sub-Saharan Africa. Additionally, RF-DTW performs robustly when transferred to untrained regions and years, with most cases showing an F1-score higher than 0.70. Our proposed method, as a combination of phenological alignment and machine learning, can be used to map soybean accurately and efficiently across different regions and years, to provide crucial information for understanding the rapid dynamics of soybean cultivation and its global-scale impacts.

Droplet Impact on the Superhydrophobic Surface with Singularities.

Reducing the contact time of an impacting droplet is highly desirable in various industrial fields including anti-icing. With the straightforward upscaling advantage, singularities on superhydrophobic surfaces can induce an annular rebound with a limited reduction in contact time. To break this limitation and further reduce contact time, this study focuses on optimizing the singularity number and arrangement. The effects of the singularity number and dimensionless spacing (l* scaled by the droplet diameter) on the dynamic and contact time characteristics of a droplet impacting the superhydrophobic surface are experimentally studied under varying Weber numbers (We). The experimental results indicate that in comparison to the single singularity, two singularities with l* < 1.0 can generate two liquid rings with four lateral liquid subunits due to the impalement at the high We region. Owing to the reduced equivalent diameter of the subunit, increasing We results in a gradually decreased contact time and accordingly breaks the limitation. However, the liquid film cannot be pierced at l* > 1.0 with a limited reduction. Considering the further reducing potential at l* < 1.0, four singularities are explored without a further reduced contact time due to the formed central liquid film. Using an additional central singularity, the central liquid film is pierced promoting its annular rebound. In consequence, five singularities significantly break the limitation in contact time, particularly a 61.7% reduction to the superhydrophobic flat surface at l* < 1.0.

The application of improved DTW algorithm in sports posture recognition

Sports posture recognition plays a crucial role in modern sports science and training. Posture recognition and analysis plays a positive role in improving sports quality and ensuring sports safety. However, existing recognition technologies still have poor recognition and accuracy in large amounts of posture data. Therefore, to further improve the performance of the existing posture recognition techniques, this study assumes that postures during movement can be effectively represented through the time series of skeletal key points, and the local similarity of these postures can be captured through the Dynamic Time Warping (DTW) algorithm. Based on this assumption, the existing DTW algorithm is improved by introducing the K-Nearest Neighbor (KNN) algorithm and combining it with Principal Component Analysis (PCA) for feature dimensionality reduction. A novel algorithmic model for postures recognition is proposed. The experimental results showed that the improved algorithm performed well in postures recognition rate and accuracy. Especially, when the k value was 5, the recognition rate reached up to 89%, and the accuracy reached 87%. Compared with the existing algorithm, the improved KNN-DTW algorithm has significant improvement in accuracy and computational efficiency. In summary, the new algorithm shows significant advantages in terms of accuracy and stability, providing a powerful tool for the analysis of athletic postures in the field of sports. Meanwhile, this research result has important application prospects in fields such as sports training, sports medicine, and virtual reality.

Investigating 10 Yr of Volcanoacoustic Activity at Tungurahua Volcano, Ecuador, Aided by Machine Learning

Abstract Here, we present the results of applying diverse data processing and machine learning tools to investigate a very large dataset obtained from single station infrasonic recordings from the last 10 yr of the most recent period of explosive activity at Tungurahua volcano, Ecuador. To increase the quality and quantity of information extracted from the large data set and enhance pattern recognition, we combined traditional techniques with more recent ones. We divided the investigation into sequential steps: detection, discrimination, cleaning, and clustering. For the detection step, we tested the classical short-term average/long-term average algorithm and an algorithm specific for explosions detection called “Volcanic INfrasound Explosions Detector Algorithm (VINEDA)” and detected 118,516 events. To clean the detected signals from potential false positives, we used supervised classification that reduced the events to 75,483, and a catalog cleaning procedure using shallow learners including support vector machines, random forests, and a single layer neural network, trained using data from a manual catalog, to a final number of 36,359 events. This led to a sixfold increase in detected explosions compared to the manual catalog. Then, we applied hierarchical clustering to a well-studied time window of activity using two independent difference metrics: dynamic time warping and waveform cross correlation and showed the insights and drawbacks from this approach. We showed that the different techniques were able to reveal repeating and striving events between selected different eruptive phases and associated them to possible changes in eruptive dynamics. Finally, to analyze the whole dataset at once we used a convolutional autoencoder network and obtained similar results to the classical clustering in a fraction of the time. We identified different families of explosions that appeared, sometimes intermittently, and revealed various potentially competing eruptive processes during the whole time period.

Unsupervised clustering of longitudinal clinical measurements in electronic health records.

Longitudinal electronic health records (EHR) can be utilized to identify patterns of disease development and progression in real-world settings. Unsupervised temporal matching algorithms are being repurposed to EHR from signal processing- and protein-sequence alignment tasks where they have shown immense promise for gaining insight into disease. The robustness of these algorithms for classifying EHR clinical data remains to be determined. Timeseries compiled from clinical measurements, such as blood pressure, have far more irregularity in sampling and missingness than the data for which these algorithms were developed, necessitating a systematic evaluation of these methods. We applied 30 state-of-the-art unsupervised machine learning algorithms to 6,912 systematically generated simulated clinical datasets across five parameters. These algorithms included eight temporal matching algorithms with fourteen partitional and eight fuzzy clustering methods. Nemenyi tests were used to determine differences in accuracy using the Adjusted Rand Index (ARI). Dynamic time warping and its lower-bound variants had the highest accuracies across all cohorts (median ARI>0.70). All 30 methods were better at discriminating classes with differences in magnitude compared to differences in trajectory shapes. Missingness impacted accuracies only when classes were different by trajectory shape. The method with the highest ARI was then used to cluster a large pediatric metabolic syndrome (MetS) cohort (N = 43,426). We identified three unique childhood BMI patterns with high average cluster consensus (>70%). The algorithm identified a cluster with consistently high BMI which had the greatest risk of MetS, consistent with prior literature (OR = 4.87, 95% CI: 3.93-6.12). While these algorithms have been shown to have similar accuracies for regular timeseries, their accuracies in clinical applications vary substantially in discriminating differences in shape and especially with moderate to high missingness (>10%). This systematic assessment also shows that the most robust algorithms tested here can derive meaningful insights from longitudinal clinical data.

Electric vehicle energy consumption prediction for unknown route types using deep neural networks by combining static and dynamic data

Accurate energy consumption prediction of electric vehicles (EVs) is crucial for drivers considering long trips. All the data should be provided beforehand to determine the energy consumption at the beginning of the trip. Although dynamic vehicle data (vehicle speed, state-of-charge, acceleration, etc.) cannot be known before the trip, factors related to the specified route (route type, elevation, average speed, weather, driving time, etc.) can be used to predict the consumed energy. These factors can be categorized as static and dynamic features, and thus, the question of how to effectively use static and dynamic data arises. This paper investigates the problem of predicting the energy consumption of an EV for a predetermined trip using a deep neural network (DNN) model that effectively uses static features along with dynamic segment features. Furthermore, we address the problem where the route types are unknown in advance. To include more information in the prediction model, we clustered the speed profiles using shape-based clustering with dynamic time warping (DTW) to predict the route type and used the cluster labels as static inputs. Real driving data collected from various drivers of a specific EV were used to train the DNN. The proposed DNN model was compared with the average energy consumption (AEC) model and five machine learning models. The results show that labels obtained from shape-based clustering improved the prediction more than feature-based cluster labels. The prediction errors were minimized with the proposed DNN model, where static features are introduced to the first and second layers twice.