Statistical Indicators Research Articles

Lube oil life prediction for heavy earth moving machinery (HEMM): A machine learning approach

Evaluating lubricant life is crucial for maintaining equipment reliability and preventing failures. Conventional methods often depend on original equipment manufacturer recommendations for lubricant changes, which may result in the premature disposal of operationally effective lubricants, leading to economic costs and degrading overall efficiency. The chemical properties of these samples are evaluated by calculating multiple parameters such as total acid number, total base number, oxidation index, soot level, and water contamination. In addition, rheological properties through viscosity index analysis and the tribological properties via friction-wear analysis are determined. In this study, an artificial neural network (ANN) (a four-layer perceptron) and an adaptive neuro-fuzzy inference system (ANFIS) are applied to predict oil conditions based on multiple calculated parameters. Performance and comparison of these advanced mathematical models are evaluated using statistical indices. Overall, the artificial intelligence (AI)-powered approach proved effective in predicting lubricant life for HEMM. Among the AI models, the ANN model demonstrated particularly strong performance, with a correlation coefficient of 0.99 compared to 0.98 for the ANFIS model. Implementing the ANN model could lead to a potential 19% reduction in current engine oil expenses, which would lower operating costs and decrease environmental impact by reducing the frequency of oil disposal.

Weak Fault Feature Extraction in Non-Gaussian Noise Interference Based on Adaptive Recombination Empirical Wavelet Transform Incorporated by Sparse Coding Shrinkage

Abstract The fast kurtogram (FK) is a very useful tool in the field of fault diagnosis, but it also contains two drawbacks. On one hand, its kurtosis indicator is susceptible to random impulse interference, leading to frequency band mis-selection. On the other hand, the frequency band division method determined by the tree filter bank may cause under or over decomposition problems. Therefore, in this paper a novel fault diagnosis theory is constructed combining the adaptive recombination empirical wavelet transform (AREWT) with the envelope spectral energy ratio (ESER). Adaptive decomposition of frequency bands is first performed using AREWT. Afterwards, the ESER is proposed as a statistical indicator to choose optimal demodulation frequency band which overcomes the effects of non-Gaussian noise interference and improves diagnostic accuracy. To further highlight the fault elements of the selected components, the adaptive sparse coding shrinkage algorithm (ASCS) is introduced to sparsely denoising sensitive components. Correspondingly, clear fault feature frequency components could be extracted from envelope spectrum. Finally, the practicability and superiority of proposed AREWT-ESER approach were fully validated through numerical simulation signal and case studies.

An Effective Robust Total Least-Squares Solution Based on “Total Residuals” for Seafloor Geodetic Control Point Positioning

Global Navigation Satellite System/Acoustic (GNSS/A) underwater positioning technology is attracting more and more attention as an important technology for building the marine Positioning, Navigation, and Timing (PNT) system. The random error of the tracking point coordinate is also an important error source that affects the accuracy of GNSS/A underwater positioning. When considering its effect on the mathematical model of GNSS/A underwater positioning, the Total Least-Squares (TLS) estimator can be used to obtain the optimal position estimate of the seafloor transponder, with weak consistency and asymptotic unbiasedness. However, the tracking point coordinates and acoustic ranging observations are inevitably contaminated by outliers because of human mistakes, failure of malfunctioning instruments, and unfavorable environmental conditions. A robust alternative needs to be introduced to suppress the adverse effect of outliers. The conventional Robust TLS (RTLS) strategy is to adopt the selection weight iteration method based on each single prediction residual. Please note that the validity of robust estimation depends on a good agreement between residuals and true errors. Unlike the Least-Squares (LS) estimation, the TLS estimation is unsuitable for residual prediction. In this contribution, we propose an effective RTLS_Eqn estimator based on “total residuals” or “equation residuals” for GNSS/A underwater positioning. This proposed robust alternative holds its robustness in both observation and structure spaces. To evaluate the statistical performance of the proposed RTLS estimator for GNSS/A underwater positioning, Monte Carlo simulation experiments are performed with different depth and error configurations under the emulational marine environment. Several statistical indicators and the average iteration time are calculated for data analysis. The experimental results show that the Root Mean Square Error (RMSE) values of the RTLS_Eqn estimator are averagely improved by 12.22% and 10.27%, compared to the existing RTLS estimation method in a shallow sea of 150 m and a deep sea of 3000 m for abnormal error situations, respectively. The proposed RTLS estimator is superior to the existing RTLS estimation method for GNSS/A underwater positioning.

Ценностные основы профессиональной подготовки будущих учителей русского языка и литературы

The purpose of the study is to identify the teacher’s priority values and ways to form them in the process of studying at the university for future teachers of Russian language and literature. The scientific novelty of the research lies in an integrated approach to solving the problem of forming the value foundations of professional training for undergraduate students in the specialty 44.03.05 Pedagogical education (with two training profiles), profiles “Russian language”, “Literature”, within the framework of the axiological direction in teaching the Russian language, which includes a wide range of opportunities to work with folklore texts, contributing to the formation of a system of traditional values and professional competencies of future teachers. Through a qualitative and quantitative analysis of the results of the survey of future language teachers, the most relevant values of a modern teacher have been identified, and the importance of each category of values has been established, taking into account the statistical indicators of students’ choice of each course of study. The research has shown that a systematic approach to the formation of the value foundations of students’ professional training and a comprehensive solution to the current problems of methodological support for future teachers of Russian language and literature are necessary conditions for the successful development of the modern educational space of the Russian Federation.

The use of a drug based on extracellular matrix biomimetics as a means of regenerative therapy during resection of the tip of the tooth root

Introduction. The operation of resection of the tip of the tooth root is one of the main dental–preserving surgical methods of dental treatment with destructive changes in periapical tissues. Despite the high efficiency and other positive aspects, there are some disadvantages of this method of treatment, among which one can single out a long period of healing of periapical tissues, which necessitates the search for remedies that contribute to solving this problem. The aim is to study the possibility of using a drug based on extracellular matrix biomimetics as a means of regenerative therapy during resection of the tip of the tooth root, to evaluate its effect on the healing time of periapical tissues and the occurrence of complications. Materials and methods. The study involved 42 patients with indications for resection of the tip of the tooth root. The patients were randomly divided into two equal groups (n = 21): the control group, which underwent a traditional operation of resection of the tip of the tooth root, and the study group, which additionally used a drug based on extracellular matrix biomimetics. The timing of complete healing of periapical tissues and the presence of complications (in the form of lack of healing) were assessed using targeted intraoral contact radiography and percussion 6 months after the operation. The collected data were statistically analyzed. The indicators of descriptive statistics included the number of objects, the average value, the mean square deviation (SKO), 95 % confidence interval (CI95 %), median, lower (Q1) and upper (Q3) quartiles, minimum and maximum values. The comparison of the healing time in the study and control groups was carried out using the nonparametric Mann – Whitney criterion for independent groups. The comparison of the groups by the number of complications was carried out using the exact Fisher criterion. Results. In the study group, healing occurred on average after 129.9 days. (COE = 8.6 days), DI95 % 126.0–133.8 days, minimum healing period – 117.0 days, maximum – 151.0 days, median – 128.0 days, lower and upper quartiles 124.0 days and 133.0 days accordingly. In the control group, the average healing time was 181.5 days. (COE = 9.0 days), DI95 % 177.4–185.6 days, minimum healing period – 167.0 days, maximum – 199.0 days, median – 181.0 days, lower quartile – 175.0, upper quartile – 188.0 days. A comparison of the groups by healing time using the nonparametric Mann – Whitney criterion for independent samples showed a statistically significant difference (p = 0.00000003). In the study group, healing was faster (on average 129.9 days) than in the control group (on average 181.5 days). A comparison of the groups by the proportion (%) of complications using the exact Fisher criterion showed that there was no statistically significant difference between the groups (p = 0.49). In the study group, healing was in 21 (100 %) patients, in the control group – in 19 (90.5 %), complications were in 2 (9.5 %) patients. Conclusion. The results obtained during the study indicate the effectiveness of using the drug based on extracellular matrix biomimetics after resection of the tip of the tooth root: its effect reduces the time for complete healing of periapical tissues and completely prevents the occurrence of complications in the form of lack of healing.

Hydrogeochemical Characterization and Source Evaluation of Groundwater Contamination in Parts of the Uddanam CKDu Area, Southern India: Integrating Statistical, Geochemical Modeling, Pollution, and Health Indices

Hydrogeochemical Characterization and Source Evaluation of Groundwater Contamination in Parts of the Uddanam CKDu Area, Southern India: Integrating Statistical, Geochemical Modeling, Pollution, and Health Indices

Region as a “Double Periphery“ (on the Example of the Russian Far East)

The study of “double peripheries” is a new area of center-periphery research, the development of which is currently hampered by several methodological difficulties, including the problems of conceptualization and empirical interpretation of the “double periphery“ as a scientific concept. Citing Russian Far East as an example, the article attempts to provide methodologically founded identification of a specific double periphery. Drawing on the results of contemporary center-periphery studies, the paper advances a working definition of a double periphery. This allows for identifying statistical indicators that are necessary to assess whether a socio-geographical object belongs to the peripheries of the named type. After that the author conducts a comparative historical analysis of quantitative data on the societal potential of the Russian Far East in 1990–2020 and the dynamics of its trade, investment and migration ties. The analysis concludes that the macroregion represents a double periphery which has developed in the post-Soviet period in relation to the NEA-3 countries (China, Japan, the Republic of Korea) and Russia. Today it is most clearly manifested in the export of goods and outbound tourism. Finally, the article outlines possible ways to stimulate development of the Far Eastern macroregion under the objective constraints imposed by its peripheral position. The implication to be made is that it is necessary to move from the selective policy of local preferences, which has been carried out since the beginning of the 2010s, to the areal planned development of the Russian Far East based on the introduction of a special administrative regime on the its territory.

Platformization in Tourism: Typology of Business Models, Evolution of Market Concentration and European Regulation Responses

The paper serves to discuss the evolution of platform business models in tourism, the dynamics of market concentration, and regulatory interventions. We present the historical timeline of digitalization in tourism distribution models, starting from the Global Distribution Systems (GDS), and continuing with Online Travel Agencies (OTAs) and their market concentration in the duopoly of Expedia Group versus Booking Holdings Group. We continue to discuss the slow, yet uncertain, breaking of the global duopoly in the travel platforms market. On the one hand, the duopoly is challenged from the “bottom up”, where Airbnb is the most successful example, riding on the wave of popularity of the ideals of the “sharing economy” and the aftermaths of the 2008 global financial crisis. On the other hand, the duopoly is challenged from the “top down”—the most powerful digital players have started to extend their operations to tourism. The second section of the paper identifies three current areas of platformization in tourism operations—distribution, promotion and HRM. We analyze available industry and public statistics indicators in Europe for each of the three dimensions. Finally, we discuss the regulatory responses to the market concentration of travel distribution platforms. Although the challenge of market concentration is a global issue, and the threats are the same in many countries, we limit the discussion of regulation to the European context. In conclusion we discuss the possibilities of future development, and point to the threat of the further acceleration and interconnection of platformization and market concentration in the travel business.

Machine learning-based estimation of crude oil-nitrogen interfacial tension

Accurate estimation of interfacial tension (IFT) between nitrogen and crude oil during nitrogen-based gas injection into oil reservoirs is imperative. The previous research works dealing with prediction of IFT of oil and nitrogen systems consider synthetic oil samples such n-alkanes. In this work, we aim to utilize eight machine learning methods of Decision Tree (DT), AdaBoost (AB), Random Forest (RF), K-nearest Neighbors (KNN), Ensemble Learning (EL), Support Vector Machine (SVM), Convolutional Neural Network (CNN) and Multilayer Perceptron Artificial Neural Network (MLP-ANN) to construct data-driven intelligent models to predict crude oil – nitrogen IFT based upon experimental data of real crude oils samples encountered in underground oil reservoirs. Several statistical indices and graphical approaches are used as accuracy performance indicators. The results show that virtually all the gathered datapoints are suitable for the purpose of model development. The sensitivity analysis indicated that pressure, temperature and crude oil API all negatively affect the IFT, with pressure being the most effective factor. The evaluation study proved that Random Forest is the most accurate developed intelligent model as it was characterized with acceptable R-squared (0.959), mean square error (1.65), average absolute relative error (6.85%) of unseen test datapoints as well as with correct trend prediction of IFT with regard to all input parameters of pressure, temperature and crude oil API. The developed model can be considered an accurate an easy-to-use tool for the prediction of crude oil/N2 IFT values for enhance oil recovery study optimization and upstream reservoir investigations.

Distribution of race and training-related musculoskeletal injuries by track and location in Thoroughbred and Arabian racehorses.

Musculoskeletal injuries (MSI) are a major concern in the horse racing industry, often leading to career-ending outcomes. Contributing factors include conformation, limb and joint defects, hoof structure, age, and hard track surfaces. This study aimed to evaluate the distribution of MSI in Thoroughbred and Arabian racehorses during racing and training, categorised by breed and track surface. Retrospective cohort study. The data from 2018 to 2023 were collected at the Turkish Jockey Club's Şirinyer Hippodrome (38°23'03.7″ N 27°09'01.0″ E), covering 553 MSI cases. The distribution of MSI, categorised by track surface and race/training, was evaluated using a 2 × 3 Pearson chi-square analysis and a post hoc Z-test. MSI occurrences during racing and training were examined based on breed and track surface by gender, age, category, location, and impact on the horses' racing career, was calculated using descriptive statistical indicators such as frequency and percentage. The majority of MSI occurred between June and September, the summer and dry period in the northern hemisphere. The 553 MSI cases included 45.3% fractures, 40.3% tendon injuries, 9.8% bone fissure. Distribution of MSI occurring during races and training according to track surface in Thoroughbred and Arabian horses had significant differences (p < 0.0001). Among Thoroughbreds, injuries were most common at 2- and 3-year-old, while in Arabians, they were most frequent at 3- and 4-year-old. The highest prevalence of MSI occurred in the forelimbs during racing, affecting 91.3% of Thoroughbreds and 92.7% of Arabians, with a particular focus on the carpal joint and its distal structures. MSI ended the racing careers of 66.3% of Thoroughbreds and 46.4% of Arabians. This study relied on routinely recorded hospital data. Fractures predominated in Thoroughbreds, particularly on dirt tracks, while tendon injuries were common in Arabians across both surfaces. Findings indicate the need to improve track maintenance and training practices to support racehorse welfare and longevity.

Methane Adsorption in Heterogeneous Potential Wells of Coal: Characterization Model and Applications.

In terms of the phenomenon of nonuniformity adsorption energy between methane and a natural heterogeneous coal surface, a heterogeneous potential well model is established in this study based on adsorption science and molecular dynamics theories. This model describes the methane adsorption positions in coal pores as a three-dimensional space composed of adsorption equipotential surfaces with varying depths of potential well, which emphasizes the heterogeneous distribution of methane adsorption potential well depths in coal and accurately describes the spatial distribution and energy states of methane molecules during methane adsorption and desorption in naturally heterogeneous coal. By taking the residual sum of squares (RSS) and Pearson correlation coefficient as indicators, the fitting accuracies of the Langmuir model and the heterogeneous potential well model for isothermal adsorption and desorption curves are compared so that the superiority of the heterogeneous potential well model in describing the adsorption and desorption of methane in natural coal is confirmed. According to this new model, a method to calculate the potential well distribution of coal by using isotherm adsorption and desorption curves was proposed. Taking the number of potential wells, the average potential well depth, and the variance of potential well depth as statistical indicators, the potential well distribution characteristics of coal in different ranks in the processes of methane adsorption and desorption under different temperatures were analyzed. In addition, the effects of temperature increase on the changes of the occupation rate of potential wells and methane desorption amount of coal with different potential well depth distributions are studied, which confirmed the necessity of evaluating the thermal recovery rate of coalbed methane based on the potential well depth distribution of coal seams.

The fractional nonlinear Schrödinger equation: Soliton turbulence, modulation instability, and extreme rogue waves.

In this paper, we undertake a systematic exploration of soliton turbulent phenomena and the emergence of extreme rogue waves within the framework of the one-dimensional fractional nonlinear Schrödinger (FNLS) equation, which appears in many fields, such as nonlinear optics, Bose-Einstein condensates, plasma physics, etc. By initiating simulations with a plane wave modulated by small noise, we scrutinized the universal regimes of non-stationary turbulence through various statistical indices. Our analysis elucidates a marked increase in the probability of rogue wave occurrences as the system evolves within a certain range of Lévy index α, which can be ascribed to the broadened modulation instability bandwidth. This heightened probability of extreme rogue waves is corroborated through multiple facets, including wave-action spectrum, fourth-order moments, and probability density functions. However, it is crucial to acknowledge that a decrease in α also results in a reduction in the propagation speed of solitons within the system. Consequently, only high-amplitude solitons with non-zero background are observed, and the occurrence of collisions that could generate higher-amplitude rogue waves is suppressed. This introduces an inverse competitive mechanism: while a lower α expands the bandwidth of modulation instability, it concurrently impairs the mobility of solitons. Our findings contribute to a deeper understanding of the mechanisms driving the formation of rogue waves in nonlinear fractional systems, offering valuable insights for future theoretical and experimental studies.

Ostracod-based transfer function shifting to a broad prospect in palaeolimnology and palaeoclimate.

The ostracod-based transfer function is a common method in palaeolimnology and palaeoclimate research, serving as an interpreter from fossil record to palaeoenvironment history. However, migration application of the transfer function remains challenging, and it is characterized by the use of the threshold of the transfer function and restriction of the understanding of palaeolimnology and palaeoclimate. Here, we constructed transfer functions for water salinity, alkalinity, and Ca conditions employing ecological information about four key species in the Yamdrok-tso basin. The prediction power of the transfer functions and their migration application in the Pumoyum Co were evaluated. The results show that the prediction power of the transfer functions was maintained according to the statistical indices of the cross-validation (R2, root mean square error, Rjack2, and root mean square error of prediction). The palaeoenvironment records constructed for the Pumoyum Co using transfer functions are in good agreement with the lake depth, mean grain size, ostracod diversity, and stalagmite δ18O records, indicating the considerable applicability of the transfer function in different ecosystems. Our study provides fresh insights for the application of transfer functions to palaeoenvironment reconstruction, enhancing the comprehension of palaeolimnology and palaeoclimate.

Enhancing air quality predictions in Chile: Integrating ARIMA and Artificial Neural Network models for Quintero and Coyhaique cities.

In this comprehensive analysis of Chile's air quality dynamics spanning 2016 to 2021, the utilization of data from the National Air Quality Information System (SINCA) and its network of monitoring stations was undertaken. Quintero, Puchuncaví, and Coyhaique were the focal points of this study, with the primary objective being the construction of predictive models for sulfur dioxide (SO2), fine particulate matter (PM2.5), and coarse particulate matter (PM10). A hybrid forecasting strategy was employed, integrating Autoregressive Integrated Moving Average (ARIMA) models with Artificial Neural Networks (ANN), incorporating external covariates such as wind speed and direction to enhance prediction accuracy. Vital monitoring stations, including Quintero, Ventanas, Coyhaique I, and Coyhaique II, played a pivotal role in data collection and model development. Emphasis on industrial and residential zones highlighted the significance of discerning pollutant origins and the influence of wind direction on concentration measurements. Geographical and climatic factors, notably in Coyhaique, revealed a seasonal stagnation effect due to topography and low winter temperatures, contributing to heightened pollution levels. Model performance underwent meticulous evaluation, utilizing metrics such as the Akaike Information Criterion (AIC), Ljung-Box statistical tests, and diverse statistical indicators. The hybrid ARIMA-ANN models demonstrated strong predictive capabilities, boasting an R2 exceeding 0.90. The outcomes underscored the imperative for tailored strategies in air quality management, recognizing the intricate interplay of environmental factors. Additionally, the adaptability and precision of neural network models were highlighted, showcasing the potential of advanced technologies in refining air quality forecasts. The findings reveal that geographical and climatic factors, especially in Coyhaique, contribute to elevated pollution levels due to seasonal stagnation and low winter temperatures. These results underscore the need for tailored air quality management strategies and highlight the potential of advanced modeling techniques to improve future air quality forecasts and deepen the understanding of environmental challenges in Chile.

Estratégias de irrigação na produtividade da água e da cultura do girassol com base em experimentos de campo e modelação

ABSTRACT Soil water balance models can be a good option for studying crop growth and yield responses under different water supply levels. The aim of this study was to evaluate the effect of different irrigation strategies on sunflower yield and water productivity using field experiments and soil water balance simulations. Two field experiments were performed. In 2018/19, four irrigation strategies (aimed at maintaining soil available water at 80, 70, 60, and 50% of total available water) were evaluated. In 2020/21 an irrigated and a rain-fed treatment were evaluated. The SIMDualKc model was calibrated and validated to simulate soil water balance and derive the dual coefficients using plant variables from field observations. Grain yield was highest in the treatment maintained at 70% of total available water in Study I, suggesting that a water replenishment reduction strategy may be an alternative to irrigation management, especially in locations where water is scarce. The statistical indicators showed good performance of the model in simulating the water available in the soil, with the coefficient of determination &gt; 0.90 and modeling efficiency above 0.86 for all cases. The study allowed the calibration of the single and basal crop coefficients of sunflower, which are important parameters for irrigation management.

In sillico study of a series of molecules derived from thioureas with breast anticancer activities of the MCF-7 strain

In the present study, a Quantitative Structure Activity Relationship (QSAR) against breast cancer of the MCF-7 strain was carried out using a series of twenty-five (25) molecules derived from Thio-ureas. The molecular descriptors were obtained after optimization of all these molecules at the B3LYP/6-31+ G (d, p) calculation level. The multiple linear regression (MLR) method was used to carry out this study. The use of this method thus made it possible to obtain a model from the molecular descriptors which are lipophilicity LogP, the bond lengths d(C=N1) and d(N2-Cphen1), the bond angle A(O-C - Cphen2) and the number of fluorine atoms F in the molecule. Furthermore, the stability, exploratory and predictive power of the model obtained was achieved by internal and external validation methods and their area of applicability was verified by the levers approach. All compounds had their leverage value lower than the critical value (h*=1.06), so all belong to the applicability domain. The results from the statistical indicators obtained from the model (R2 =0.976; RMCE=0.073; F= 89.502), made it possible to affirm that this model is acceptable, robust with good predictive power. Lipophilicity (LogP) turned out to be the priority descriptor in predicting the breast anticancer activity of the MCF-7 strain of our QSAR model.

Comparative analysis of statistical indicators of cervical cancer and their relations to the income of the population in the Arctic Zone of Russia in 2016-2022

Comparative analysis of statistical indicators of cervical cancer and their relations to the income of the population in the Arctic Zone of Russia in 2016-2022

Improvement of the Budget Forecasting System in the Kyrgyz Republic

The purpose of this study is to develop and propose measures to improve the budget forecasting system in Central Asia to enhance the accuracy, reliability, and adaptability of budget forecasts. The study involved a comprehensive analysis of data covering various aspects of budget forecasting in Central Asian countries. This analysis included the collection of extensive data, including statistical indicators, on budget forecasts, and factors influencing economic stability in the region. The research results underscored the importance of budget forecasting as a tool for strategic financial planning based on systemic analysis and the use of advanced technologies. The concept of budget forecasting was highlighted as a systematic analysis aimed at predicting the financial performance of an organization over a specific period, including assessing expected income and expenses considering various factors influencing the financial situation. The study also examined the key functions of budget forecasting, including financial resource planning, optimization, and financial stability control. Special attention was paid to analysing the impact of external factors, such as economic uncertainty, using statistical methods and scenario analysis. The study also emphasized the importance of modern technologies, including machine learning and big data analysis, in improving budget forecasting processes. Overall, the research findings present important insights for practical application and further research in the field of financial management and budget planning. The findings have the potential to be used in shaping policies and reforms aimed at sustainable development and efficient utilization of public finances in the region.

Reducing the Dimensionality of Data Arrays Using Multi-Layer Autoencoders in the Task of Classifying Mobile Applications

The problem of reducing the dimension of the initial data arrays to improve the efficiency of mobile application traffic processing is considered. The relevance of the study is due to the need to optimize the volume of transmitted and stored data when working in conditions of limited computing resources, as well as to increase the speed and quality of analytical operations. To solve this problem, multi-layer autoencoders are used, capable of forming compact representations of the source data with minimal losses in their informativeness. The approach is based on the idea of training neural network models that extract the most significant features from the source arrays and are able to restore them with a given level of accuracy. Methods used. During the experiments, various architectures of multilayer autocoders were used, differing in the number of layers and dimensions of hidden representations. The research was conducted on real data sets collected from mobile applications with a wide range of functionality. The analysis was carried out by varying the internal parameters of the networks and evaluating the results through an integral statistical indicator reflecting the degree of compression. This indicator allows you to identify how much the spread of attributes changes when passing data through the autoencoder. Results. To evaluate the filtering properties of multilayer autoencoders, an integral compression indicator is proposed that characterizes the change in the spread of attributes of mobile applications when passing them through an autoencoder of a given structure. The indicator is calculated as the ratio of the standard deviation of the attributes at the input and at the output, which allows you to assess the degree of data compression and the degree of information preservation after processing. It is shown that an increase in the integral compression index indicates a more significant compression of the initial data. It was found that filtering is practically independent of the type of application and lies within 10-20 % for three-layer autoencoders, whereas for five-layer auto-encoders, preference is given to encoders with a minimum dimension of the inner layer. The main novelty of the work lies in the development of an integral statistical indicator that not only reflects the degree of compression of mobile application data, but also takes into account the preservation of the original information structure. Unlike existing approaches, this indicator allows for a systematic comparison of various architectures of autoencoders, taking into account not only the reduction in dimension, but also the quality of recovery of the original information. This creates the basis for a more objective assessment of the effectiveness of multilayer autoencoders in specific application conditions. Practical significance. The proposed methodology may be useful for developers and researchers working on optimizing systems for collecting, storing and processing mobile application data. In conditions of limited computing resources, which are typical for mobile devices and embedded systems, the use of multilayer autoencoders aimed at achieving a given balance between compression and preservation of information provides a significant reduction in the volume of transmitted data. The results of the study can be implemented into existing analytical platforms, monitoring systems and classification of mobile applications.

Using classification methods in forecasting the level of geomagnetic field disturbance based on the &lt;i&gt;Kp&lt;i&gt;-Index

The paper explores the possibilities of using data classification methods when forecasting time series of the geomagnetic Kp-index by machine learning methods. To classify categories of the Kp-index based on the degree of disturbance, linear and logistic regression, random forest, gradient boosting on top of decision trees, and artificial neural networks of various architectures are used. The results of these methods are compared with a trivial inertial forecast (the statistical indicators of which for problems of this type are always high) at horizons from 3 h to 1 day in 3-h increments. The problem of choosing a cross-validation scheme for selecting the model hyperparameters, ways to overcome the imbalance of categories, the relative importance of input features, as well as the dependence of the results on the test sample (beginning of the 25th solar activity cycle) on inclusion in the training sample of data from the 23rd and 24th cycles or only the 24th cycles are studied. Based on the results, conclusions are drawn about the preferred methods for classifying values of the Kp-index based on the level of geomagnetic disturbance. Ways for further research and possible improvement of the classification quality are outlined, including for determining the characteristic hidden states of Earth’s magnetosphere as a dynamic system in order to improve the quality of forecasting geomagnetic indices.