Prefiltering Of Data Research Articles

Tourism and uncertainty: a machine learning approach

ABSTRACT In this paper, we attempt to create a unique forecasting model to forecast out-of-sample the tourism demand in 24 European Union countries. The initial dataset included 34 relevant variables of annual frequency that span the period from 2010 to 2020 for 40 countries. A data prefiltering process resulted in a final set of 17 relevant variables for 24 countries. Additionally, in the effort to investigate the impact of uncertainty on international tourism, apart from the traditional factors that affect tourism, we also include variables that measure various forms of uncertainty: we use the World Pandemic Uncertainty (WPU) Index, the Global CBOE Volatility Index, the Political Globalisation Index, the Economic Globalisation Index, and the Political Stability Index. In the empirical part of our research, we employ and compare in terms of their forecasting accuracy a set of six state-of-the-art machine learning algorithms, the Support Vector Regression with both a linear and an RBF kernel, the Random Forests, the Decision Trees, the KNN, and gradient-boosting trees. The results show that the Gradient-Boosting Trees algorithm outperforms the other five models providing the most accurate forecasts with a MAPE of 0.10% and 1.36% in the training and the out-of-sample tests, respectively.

Evaluating cell type deconvolution in FFPE breast tissue: application to benign breast disease

Abstract Transcriptome profiling using RNA sequencing (RNA-seq) of bulk formalin-fixed paraffin-embedded (FFPE) tissue blocks is a standard method in biomedical research. However, when used on tissues with diverse cell type compositions, it yields averaged gene expression profiles, complicating biomarker identification due to variations in cell proportions. To address the need for optimized strategies for defining individual cell type compositions from bulk FFPE samples, we constructed single-cell RNA-seq reference data for breast tissue and tested cell type deconvolution methods. Initial simulation experiments showed similar performances across multiple commonly used deconvolution methods. However, the introduction of FFPE artifacts significantly impacted their performances, with a root mean squared error (RMSE) ranging between 0.04 and 0.17. Scaden, a deep learning-based method, consistently outperformed the others, demonstrating robustness against FFPE artifacts. Testing these methods on our 62-sample RNA-seq benign breast disease cohort in which cell type composition was estimated using digital pathology approaches, we found that pre-filtering of the reference data enhanced the accuracy of most methods, realizing up to a 32% reduction in RMSE. To support further research efforts in this domain, we introduce SCdeconR, an R package designed for streamlined cell type deconvolution assessments and downstream analyses.

Holomics - a user-friendly R shiny application for multi-omics data integration and analysis

An organism’s observable traits, or phenotype, result from intricate interactions among genes, proteins, metabolites and the environment. External factors, such as associated microorganisms, along with biotic and abiotic stressors, can significantly impact this complex biological system, influencing processes like growth, development and productivity. A comprehensive analysis of the entire biological system and its interactions is thus crucial to identify key components that support adaptation to stressors and to discover biomarkers applicable in breeding programs or disease diagnostics. Since the genomics era, several other ’omics’ disciplines have emerged, and recent advances in high-throughput technologies have facilitated the generation of additional omics datasets. While traditionally analyzed individually, the last decade has seen an increase in multi-omics data integration and analysis strategies aimed at achieving a holistic understanding of interactions across different biological layers. Despite these advances, the analysis of multi-omics data is still challenging due to their scale, complexity, high dimensionality and multimodality. To address these challenges, a number of analytical tools and strategies have been developed, including clustering and differential equations, which require advanced knowledge in bioinformatics and statistics. Therefore, this study recognizes the need for user-friendly tools by introducing Holomics, an accessible and easy-to-use R shiny application with multi-omics functions tailored for scientists with limited bioinformatics knowledge. Holomics provides a well-defined workflow, starting with the upload and pre-filtering of single-omics data, which are then further refined by single-omics analysis focusing on key features. Subsequently, these reduced datasets are subjected to multi-omics analyses to unveil correlations between 2-n datasets. This paper concludes with a real-world case study where microbiomics, transcriptomics and metabolomics data from previous studies that elucidate factors associated with improved sugar beet storability are integrated using Holomics. The results are discussed in the context of the biological background, underscoring the importance of multi-omics insights. This example not only highlights the versatility of Holomics in handling different types of omics data, but also validates its consistency by reproducing findings from preceding single-omics studies.

Energy justice and gaps in sustainable development: A convergence testing and clustering study

The purpose of the study is to examine the dynamics of carbon emissions inequality and its effect on energy justice, whether carbon footprints have increased or decreased across countries, sectors, and households over time. We use convergence analysis, which is a methodology that effectively reveals carbon footprint dynamics. Our database contained carbon footprint data from different sectors across 180 countries between 1980 and 2020. We account for the non-linearity of the carbon footprint convergence or divergence process over time and individual heterogeneity across countries and sectors (data pre-filtering, carbon footprint convergence testing, and club clustering). Our results demonstrate a significant disparity in carbon footprints across various sectors and countries, especially among the top 1 % of earners. The wealthiest individuals worldwide have not converged towards a common trajectory, indicating high levels of carbon footprint inequality. This investigation assumes that the carbon footprint convergence or divergence process is not linear over time, allowing for fluctuations in carbon emissions inequality. Future research may explore the factors contributing to this non-linearity and its implications for energy justice. Our findings highlight global cooperation's importance in promoting sustainable production practices and a fair distribution of energy-related responsibilities. Energy production and consumption burdens are not equitably distributed, posing a significant challenge to achieving energy justice. This study provides a comprehensive understanding of the dynamics of carbon emissions inequality and its impact on energy justice. It employs a unique methodology and a multidimensional database, providing new insights into the carbon footprint disparity across sectors and countries.

A Novel Gait Phase Recognition Method Based on DPF-LSTM-CNN Using Wearable Inertial Sensors.

Gait phase recognition is of great importance in the development of rehabilitation devices. The advantages of Long Short-Term Memory (LSTM) and Convolutional Neural Network (CNN) are combined (LSTM-CNN) in this paper, then a gait phase recognition method based on LSTM-CNN neural network model is proposed. In the LSTM-CNN model, the LSTM layer is used to process temporal sequences and the CNN layer is used to extract features A wireless sensor system including six inertial measurement units (IMU) fixed on the six positions of the lower limbs was developed. The difference in the gait recognition performance of the LSTM-CNN model was estimated using different groups of input data collected by seven different IMU grouping methods. Four phases in a complete gait were considered in this paper including the supporting phase with the right hill strike (SU-RHS), left leg swimming phase (SW-L), the supporting phase with the left hill strike (SU-LHS), and right leg swimming phase (SW-R). The results show that the best performance of the model in gait recognition appeared based on the group of data from all the six IMUs, with the recognition precision and macro-F1 unto 95.03% and 95.29%, respectively. At the same time, the best phase recognition accuracy for SU-RHS and SW-R appeared and up to 96.49% and 95.64%, respectively. The results also showed the best phase recognition accuracy (97.22%) for SW-L was acquired based on the group of data from four IMUs located at the left and right thighs and shanks. Comparably, the best phase recognition accuracy (97.86%) for SU-LHS was acquired based on the group of data from four IMUs located at left and right shanks and feet. Ulteriorly, a novel gait recognition method based on Data Pre-Filtering Long Short-Term Memory and Convolutional Neural Network (DPF-LSTM-CNN) model was proposed and its performance for gait phase recognition was evaluated. The experiment results showed that the recognition accuracy reached 97.21%, which was the highest compared to Deep convolutional neural networks (DCNN) and CNN-LSTM.

A Novel Model Predictive Controller for Distributed Generation in Isolated Microgrids—Part I: Development and Parameterization of the Data-Driven Predictive Model

In a low-voltage islanded microgrid, the distribution line impedance and relatively large power angle may lead to active and reactive power coupling during voltage and frequency control actions, which cause errors for conventional droop control at interfacing inverters of distributed generation (DG) units. To overcome this issue, in this two-part article, a novel model predictive controller for DGs is proposed to regulate voltage and frequency at the point of common coupling in isolated microgrids. In Part 1 of the article, a data-driven predictive model is developed and parameterized by system identification approach using nonlinear least square (NLS) method. Four NLS optimization algorithms (Gauss–Newton (GN), adaptive GN, Levenberg–Marquardt, and trust region reflective)] are evaluated and GN is chosen as the best performing algorithm. To initialize the iteration for NLS, the Backcast technique is chosen after comparing with Zero and Estimate techniques. Two optimization methods, “simulation” and “prediction” focus, are considered; and the prediction focus shows higher prediction accuracy and faster convergence, and the ability to avoid the necessity of data prefiltering by introducing a built-in weighted filter in the objective function. The polynomial input–output Box–Jenkins model is chosen as the model structure. The model incorporates distribution line parameters into the control algorithm and allows a wider variation of power angle without initiating nonlinearity. Therefore, it can substantially reduce the controller size and complexity, and widen the controller's operational range. The model is further implemented in the proposed model predictive controller in Part 2 of the article.

IEEE Transactions on Sustainable Computing Special Issue on Sustainability of Fog/Edge Computing Systems

The papers in this special section focus on sustainability of edge computing systems. Edge computing is an emerging architectural and technical approach aimed at addressing various shortcomings in traditional cloud computing paradigms and responding to today’s constantly increasing data-demanding services such as Internet-of-Things, 5G embedded artificial intelligence and smart cities. In Fog/Edge Computing, nodes at the edge of a network are equipped with processing, storage, networking, etc. capabilities to take over several tasks that were used to be sent to cloud services. Pre-filtering and aggregation of data as well as online processing and actuation are sample procedures envisaged/dedicated to fog/ edge nodes.

Investigating the Short-Term Effects of Cold Stress on Metabolite Responses and Metabolic Pathways in Inner-Mongolia Sanhe Cattle.

Simple SummaryCold stress is a major environmental stressor affecting cattle performance in temperate regions, which causes impaired welfare and economic losses to cattle producers. The identification of biological mechanisms associated with cold stress response is paramount for developing effective mitigation strategies, such as genomic selection. In this study, we assessed the short-term effects of hyper-cold stress on metabolite responses and metabolic pathways in the serum of Inner-Mongolia Sanhe cattle. Moreover, 19 differential metabolites were found, mainly involved in amino acid metabolism. A further integration of metabolome results and gene expression highlighted the regulation of metabolic changes and related pathways in severe cold exposure, such as “aminoacyl-tRNA biosynthesis” and “valine, leucine, and isoleucine degradation”. In summary, we presented new insights on the short-term effects of severe cold stress as well as metabolites and metabolic pathways associated with cold stress response in Inner-Mongolia Sanhe cattle.Inner-Mongolia Sanhe cattle are well-adapted to low-temperature conditions, but the metabolic mechanisms underlying their climatic resilience are still unknown. Based on the 1H Nuclear Magnetic Resonance platform, 41 metabolites were identified and quantified in the serum of 10 heifers under thermal neutrality (5 °C), and subsequent exposure to hyper-cold temperature (−32 °C) for 3 h. Subsequently, 28 metabolites were pre-filtrated, and they provided better performance in multivariate analysis than that of using 41 metabolites. This indicated the need for pre-filtering of the metabolome data in a paired experimental design. In response to the cold exposure challenge, 19 metabolites associated with cold stress response were identified, mainly enriched in “aminoacyl-tRNA biosynthesis” and “valine, leucine, and isoleucine degradation”. A further integration of metabolome and gene expression highlighted the functional roles of the DLD (dihydrolipoamide dehydrogenase), WARS (tryptophanyl-tRNA synthetase), and RARS (arginyl-tRNA synthetase) genes in metabolic pathways of valine and leucine. Furthermore, the essential regulations of SLC30A6 (solute carrier family 30 (zinc transporter), member 6) in metabolic transportation for propionate, acetate, valine, and leucine under severe cold exposure were observed. Our findings presented a comprehensive characterization of the serum metabolome of Inner-Mongolia Sanhe cattle, and contributed to a better understanding of the crucial roles of regulations in metabolites and metabolic pathways during cold stress events in cattle.

Rescaled variance tests for seasonal stationarity

Abstract This paper introduces rescaled variance [V/S] tests for seasonal stationarity. The V/S statistic is designed by Giraitis, L., P. Kokoszka, R. Leipus, and G. Teyssière. 2003. “Rescaled Variance and Related Tests for Long Memory in Volatility and Levels.” Journal of Econometrics 112: 265–94 to be the mean corrected versions of the KPSS statistic. In the seasonal context, Canova, F., and B. E. Hansen. 1995. “Are Seasonal Patterns Constant over Time? A Test for Seasonal Stability.” Journal of Business & Economic Statistics 13: 237–52 present the seasonal generalization of the KPSS statistic. In this regard, I aim to strengthen the work of Canova, F., and B. E. Hansen. 1995. “Are Seasonal Patterns Constant over Time? A Test for Seasonal Stability.” Journal of Business & Economic Statistics 13: 237–52 [CH] by mean correction in the seasonal framework. I obtain the asymptotic distributions of the seasonal V/S tests. The V/S tests enjoy better power performance than the CH tests while exhibiting similiar size performance. Furthermore, by data pre-filtering, I propose robustified versions of the V/S statistics to eliminate the unattended unit root problem observed in the CH tests.

Time Series Forecasting for Wind Energy Systems Based on High Order Neural Networks

Wind energy is one of the most promising alternatives as energy sources; however, to obtain the best results, producers need to forecast the wind speed, generated power and energy price in order to provide the appropriate tools for optimal operation, planning, control and marketing both for isolated wind systems and for those that are interconnected to a main distribution network. For the present work, a novel methodology is proposed for the forecasting of time series in wind energy systems; it consists of a high-order neural network that is trained on-line by the extended Kalman filter algorithm. Unlike most modern artificial intelligence methods of forecasting, which are based on hybridizations, data pre-filtering or deep learning methods, the proposed method is based on the simplicity of implementation, low computational complexity and real-time operation to produce 15-step-ahead forecasting in a time series of wind speed, generated power and energy price. The proposed scheme has been evaluated using real data from open access repositories of wind farms. The results show that an on-line training of the neural network produces high precision, without the need for any other information beyond a few past observations.

Occluded Image Function: A Novel Measure for Evaluating Machine Learning Classifiers for Biometrics

A novel method is presented for evaluating the efficacy of object recognition algorithms on occluded images, called the occluded image function (OIF). The OIF describes system behavior in occluded environments and thus gives qualitative insight into their mechanisms; derivative metrics from OIF can also be used to quantitatively compare classifiers. To showcase the utility of the OIF, an experiment is performed by obstructing optical gait images from two biped robot models and using four binary machine learning classifiers to distinguish between them. The OIF diagrams are created from each experiment, and the resulting insights about the classifiers are discussed. Using the OIF, it is shown that the primitive classifiers can sometimes perform better under occlusion conditions, possibly due to pre-filtering of gait data by uniform occlusions. This result serves to demonstrate that the OIF is a useful tool for classifier evaluation.

Razorback, an Open Source Python Library for Robust Processing of Magnetotelluric Data

Magnetotellurics (MT) is a geophysical method that investigates the relationships among the different components of the natural electromagnetic field related to the geoelectric structure of the subsurface. Data can be contaminated by anthropic noise sources and suffer from transient noise to signal variations. Since the 80s, robust processing methods have been introduced to minimize the impact of noise on sounding quality. This paper presents Razorback, an open source Python library, implemented to handle, manipulate, and combine time series of synchronous data. This modular library allows users to plug in data prefilters and includes both M-estimator and bounded influence techniques, as well as a two-stage multiple remote reference. Validation of this library is performed on a real data set by comparing the results with those of an existing code. In contrast to standalone codes, the developed library allows for the design of complex and specific processing procedures. As examples, Razorback is used to perform (i) continuous time lapse processing and (ii) processing of one site in a peri-urban context. In the latter case, we have tested all possible combinations of remote reference stations in an MT array. Our phase tensor analysis shows that the bounded influence outperforms the M-estimator in reducing the impacts of man-made electromagnetic noise on magnetotelluric soundings. The Razorback library is available at https://github.com/BRGM/razorback. Jupyter notebooks for data handling and MT robust processing are available at https://github.com/BRGM/razorback/blob/doc/docs/source/tutorials/.

GPU-based efficient join algorithms on Hadoop

The growing data has brought tremendous pressure for query processing and storage, so there are many studies that focus on using GPU to accelerate join operation, which is one of the most important operations in modern database systems. However, existing GPU acceleration join operation researches are not very suitable for the join operation on big data. Based on this, this paper speeds up nested loop join, hash join and theta join, combining Hadoop with GPU, which is also the first to use GPU to accelerate theta join. At the same time, after the data pre-filtering and pre-processing, using Map-Reduce and HDFS in Hadoop proposed in this paper, the larger data table can be handled, compared to existing GPU acceleration methods. Also with Map-Reduce in Hadoop, the algorithm proposed in this paper can estimate the number of results more accurately and allocate the appropriate storage space without unnecessary costs, making it more efficient. The rigorous experiments show that the proposed method can obtain 1.5 to 2 times the speedup, compared to the traditional GPU acceleration equi join algorithm. And in the synthetic data set, the GPU version of the proposed method can get 1.3 to 2 times the speedup, compared to CPU version.

Wavelets base Automatic Extraction Technology for Vibration Signal Characteristics of Hybrid Electric Vehicle Engine

Traditional automatic characteristic extraction technology of engine vibration signals for hybrid electric vehicles (HEV) only focuses on the analysis of engine vibration signals in time domain and frequency domain. Single time domain analysis or single frequency domain analysis cannot accurately analyse the vibration signals, while both time domain analysis and frequency domain analysis have cross-analysis. As a result, the analysis results are repetitive and conflicting, which makes it difficult to extract the characteristics of engine vibration signals. The final extraction accuracy is not high and the extraction efficiency is low. For this reason, an automatic characteristic extraction technology of HEV engine vibration signal based on wavelet packet energy analysis is proposed. Firstly, the mechanical vibration of engine is converted into corresponding voltage and current signals by various sensors and then converted into digital signals by A/D (analog/digital) conditioner. The data of vibration signals are often mixed with various noises, which have a great impact on the final analysis of vibration signals. Data interception and pre-filtering are adopted. Wave, zero-mean, elimination of trend term and elimination of staggered points are used to pre-treat the vibration signals with mixed noise. Short-Time Fourier Transform (STFT) algorithm is introduced to analyse the pre-processed engine vibration signals and the fundamental properties of the non-stationary vibration signals in actual operation of the engine are obtained. The energy distribution of the analysed engine vibration signal is calculated by the wavelet packet energy analysis method. The calculated parameters of the energy distribution of the wavelet packet are taken as the characteristic parameters of the vibration signal. The vibration signal characteristics of the engine are automatically extracted. The experiment is carried out in the form of comparison with the traditional method. The experimental results show that the time-frequency joint analysis applied in the proposed technology can accurately analyse the essential characteristics of the engine vibration signal of HEV. The wavelet packet energy analysis method can ensure the extraction accuracy of the engine vibration signal characteristics.

Classification of Sleep Apnea Severity by Electrocardiogram Monitoring Using a Novel Wearable Device.

Sleep apnea (SA) is a prevalent disorder diagnosed by polysomnography (PSG) based on the number of apnea–hypopnea events per hour of sleep (apnea–hypopnea index, AHI). PSG is expensive and technically complex; therefore, its use is rather limited to the initial diagnostic phase and simpler devices are required for long-term follow-up. The validity of single-parameter wearable devices for the assessment of sleep apnea severity is still debated. In this context, a wearable electrocardiogram (ECG) acquisition system (ECG belt) was developed and its suitability for the classification of sleep apnea severity was investigated using heart rate variability analysis with or without data pre-filtering. Several classification algorithms were compared and support vector machine was preferred due to its simplicity and overall performance. Whole-night ECG signals from 241 patients with a suspicion of sleep apnea were recorded using both the ECG belt and patched ECG during PSG recordings. 65% of patients had an obstructive sleep apnea and the median AHI was 21 [IQR: 7–40] . The classification accuracy obtained from the ECG belt (accuracy: 72%, sensitivity: 70%, specificity: 74%) was comparable to the patched ECG (accuracy: 74%, sensitivity: 88%, specificity: 61%). The highest classification accuracy was obtained for the discrimination between individuals with no or mild SA vs. moderate to severe SA. In conclusion, the ECG belt provided signals comparable to patched ECG and could be used for the assessment of sleep apnea severity, especially during follow-up.

THE METHOD OF CONSTRUCTING RECOMMENDATIONS ONLINE ON THE TEMPORAL DYNAMICS OF USER INTERESTS USING MULTILAYER GRAPH

The problem of the online construction of a rating list of objects in the recommender system is considered. A method for constructing recommendations online using the presentation of input data in the form of a multi-layer graph based on changes in user interests over time is proposed. The method is used for constructing recommendations in a situation with implicit feedback from the user. Input data are represented by a sequence of user choice records with a time stamp for each choice. The method includes the phases of pre-filtering of data and building recommendations by collaborative filtering of selected data. At pre-filtering of the input data, the subset of data is split into a sequence of fixed-length non-overlapping time intervals. Users with similar interests and records with objects of interest to these users are selected on a finite continuous subset of time intervals. In the second phase, the pre-filtered subset of data is used, which allows reducing the computational costs of generating recommendations. The method allows increasing the efficiency of building a rating list offered to the target user by taking into account changes in the interests of the user over time.

THE PROCEDURES OF VISUAL ANALYSIS FOR MULTIDIMENSIONAL DATA VOLUMES

Abstract. The paper is devoted to problems of visual analysis of multidimensional data sets using an approach based on the construction of elastic maps. This approach is quite suitable for processing and visualizing of multidimensional datasets. The elastic maps are used as the methods of original data points mapping to enclosed manifolds having less dimensionality. Diminishing the elasticity parameters one can design map surface which approximates the multidimensional dataset in question much better. Then the points of dataset in question are projected to the map. The extension of designed map to a flat plane allows one to get an insight about the structure of multidimensional dataset. The paper presents the results of applying elastic maps for visual analysis of multidimensional data sets of medical origin. Previously developed data processing procedures are applied to improve the results obtained - pre-filtering of data, removal of separated clusters (flotation), quasi-Zoom.

Modal parameter estimation for a wetted plate under flow excitation: A challenging case in using POD

In this paper some issues related to the use of the Proper Orthogonal Decomposition (POD) for the identification of the modal parameters of 2D structures are highlighted with reference to the experimental identification of the vibration modes of a plate excited on one side by a turbulent boundary layer. The poor convergence of the POD modes to the linear normal modes is explained analytically and provides also the theoretical framework to avoid these issues by applying band-pass filtering (BP-POD). The application of the POD with data pre-filtering has allowed for the evaluation of the modal damping from the proper orthogonal coordinates by using the Random Decrement Technique to obtain the corresponding free-decay function. Though the emphasis is about methodological aspects related to the POD, as a byproduct of the performed analysis some more physical insights concerning the damping sensitivity and the hydrodynamic added mass are finally carried out.

Validating team communication data using a transmission-duration threshold and voice activity detection algorithm.

The processes underlying team effectiveness can be understood by analyzing the temporal dynamics of team communication sequences. The results of such analyses have shown that the complexity of team communication is associated with team performance on task-related variables, and hence communication complexity statistics have been proposed for use as measures for real-time feedback on team performance. In two analyses of historical team communication sequences, we found that filtering via use of a transmission-duration threshold and voice activity detection algorithm resulted in significant changes in complexity relative to not filtering the data or using a transmission-duration filter alone. The use of these filtering techniques showed significant effects on the complexity of communication sequences in both a laboratory-based experiment, with participants with little experience with voice communication protocols, and in a mission simulation with trained military operators. There was also a significant non-linear relationship between the complexity of communication sequences and task performance. However, an analysis of the impact of the changes in communication dynamics gained through filtering did not demonstrate that the changed temporal dynamics of filtered data better explained team performance. It is concluded that pre-filtering of invalid communication data should be included during the data cleaning stage of statistical analysis as a matter of good scientific practice. Furthermore, such use of filtering will ensure that inferences made about the relationship between the complexity of communication between team members and their performance are not confounded by the presence of invalid communication events.

Meteor Observations at Kazan Federal University (Russia)

SummaryThis poster paper described initial results of recent meteor observations in Kazan obtained with a new meteor radar, SKiYMET. Significant improvements in the number of registrations are being recorded, enabling better statistics.Radar observations of meteors have been carried out and archived since the mid-20th Century; at Kazan University (56°N, 49°E) they commenced in 1956, and have continued with only a few interruptions. A quasi-tomographic method (developed in-house in 2000) has been used to determine the coordinates of radiants of meteor showers, and to derive the orbital elements of small showers (microshowers).The observations provide valuable information about the distribution of meteoric matter near the Earth’s orbit. That information supports studies of stratospheric temperatures. A new radar, SKiYMET, was deployed in 2015 and measured parameters such as meteor speeds, but in 2016 its software was supplemented by an in-house package which performed more efficient pre-filtering of the primary data to eliminate non-meteor reflections from various sources. The new filtering algorithm has reduced substantially the threshold S/N ratio for detecting meteors (so fainter meteors can now be recorded), while the package also enables meteor velocities to be calculated much more efficiently than with SKiYMET, improving significantly the quality and statistical indicators of the processed data. By detecting a much larger number of meteors, it enables more detailed studies of the distribution of meteor velocities per day and per season.Daily meteor counts for 2016–2017, extracted from the observations by the KFU programme, yielded significantly higher rates than with SKiYMET. These results support much better statistics, and also enable much finer details to be discerned. In particular, plots of meteor rates against speed are found to exhibit two maxima, one near 30 km s−1 and one near 55 km s−1. This bifurcation in the relationship can also be discerned in the seasonal dependences of the speed distribution.Improvements in the measurements of the speeds of meteor particles entering the atmosphere reveal an increase in the numbers with higher speeds (> 50 km s−1). The distributions of speeds and heights show a rather strong dependence of the meteor speed upon altitude; as predicted by meteor physics, those with higher speeds begin to burn at higher altitudes. A local increase in the number of meteors that have speeds of about 55 km s−1 has also been reported in the literature. The increase could be associated with heterogeneities in the distribution of meteoric matter in the vicinity of the Earth’s orbit.