Definition of a composite [1] describes an ideal composite material with perfect structure. In real composite materials, structure is usually imperfect – composites contain various types of defects [2, 3–5], especially as the casted composites are of concern. The reason for this is a specific structure of castings, related to course of the manufacturing process. In case of metal matrix composite castings, especially regarding these manufactured by saturation, there is no classification of these defects [2, 4]. Classification of defects in castings of classic materials (cast iron, cast steel, non-ferrous alloys) is insufficient and requires completion of specific defects of mentioned materials. This problem (noted during manufacturing metal matrix composite castings with saturated reinforcement in Institute of Basic Technical Sciences of Maritime University Szczecin) has become a reason of starting work aimed at creating such classification. As a result, this paper was prepared. It can contribute to improvement of quality of studied materials and, as a consequence, improve the environment protection level.

Efforts of the scientific community led to the development of multiple screening approaches for COVID-19 that rely on machine learning methods. However, there is a lack of works showing how to tune the classification models used for such a task and what the tuning effect is in terms of various classification quality measures. Understanding the impact of classifier tuning on the results obtained will allow the users to apply the provided tools consciously. Therefore, using a given screening test they will be able to choose the threshold value characterising the classifier that gives, for example, an acceptable balance between sensitivity and specificity. The presented work introduces the optimisation approach and the resulting classifiers obtained for various quality threshold assumptions. As a result of the research, an online service was created that makes the obtained models available and enables the verification of various solutions for different threshold values on new data.

The work proposes a new method for vehicle classification, which allows treating vehicles uniformly at the stage of defining the vehicle classes, as well as during the classification itself and the assessment of its correctness. The sole source of information about a vehicle is its magnetic signature normalised with respect to the amplitude and duration. The proposed method allows defining a large number (even several thousand) of classes comprising vehicles whose magnetic signatures are similar according to the assumed criterion with precisely determined degree of similarity. The decision about the degree of similarity and, consequently, about the number of classes, is taken by a user depending on the classification purpose. An additional advantage of the proposed solution is the automated defining of vehicle classes for the given degree of similarity between signatures determined by a user. Thus the human factor, which plays a significant role in currently used methods, has been removed from the classification process at the stage of defining vehicle classes. The efficiency of the proposed approach to the vehicle classification problem was demonstrated on the basis of a large set of experimental data.

Traffic classification is an important tool for network management. It reveals the source of observed network traffic and has many potential applications e.g. in Quality of Service, network security and traffic visualization. In the last decade, traffic classification evolved quickly due to the raise of peer-to-peer traffic. Nowadays, researchers still find new methods in order to withstand the rapid changes of the Internet. In this paper, we review 13 publications on traffic classification and related topics that were published during 2009-2012. We show diversity in recent algorithms and we highlight possible directions for the future research on traffic classification: relevance of multi-level classification, importance of experimental validation, and the need for common traffic datasets.

The paper presents the results of a study of cyanobacteria and green algae assemblages occurring in various tundra types determined on the basis of mosses and vascular plants and habitat conditions. The research was carried out during summer in the years 2009–2013 on the north sea−coast of Hornsund fjord (West Spitsbergen, Svalbard Archipelago). 58 sites were studied in various tundra types differing in composition of vascular plants, mosses and in trophy and humidity. 141 cyanobacteria and green algae were noted in the research area in total. Cyanobacteria and green algae flora is a significant element of many tundra types and sometimes even dominate there. Despite its importance, it has not been hitherto taken into account in the description and classification of tundra. The aim of the present study was to demonstrate the legitimacy of using phycoflora in supplementing the descriptions of hitherto described tundra and distinguishing new tundra types. Numeric hierarchical−accumulative classification (MVSP 3.1 software) methods were used to analyze the cyanobacterial and algal assemblages and their co−relations with particular tundra types. The analysis determined dominant and distinctive species in the communities in concordance with ecologically diverse types of tundra. The results show the importance of these organisms in the composition of the vegetation of tundra types and their role in the ecosystems of this part of the Arctic.

This study addresses the problem of magnetic field emission produced by the laptop computers. Although, the magnetic field is spread over the entire frequency spectrum, the most dangerous part of it to the laptop users is the frequency range from 50 to 500 Hz, commonly called the extremely low frequency magnetic field. In this frequency region the magnetic field is characterized by high peak values. To examine the influence of laptop’s magnetic field emission in the office, a specific experiment is proposed. It includes the measurement of the magnetic field at six laptop’s positions, which are in close contact to its user. The results obtained from ten different laptop computers show the extremely high emission at some positions, which are dependent on the power dissipation or bad ergonomics. Eventually, the experiment extracts these dangerous positions of magnetic field emission and suggests possible solutions.

The article presents the results of research, the aim of which was to determine the qualitative and quantitative structure of the causes of accidents that were a result of falling from scaffolding. An original methodology for the classification of accidents with regards to their causes was developed and was based on cluster analysis. An example of using the proposed methodology is provided. 187 post-accident protocols of occupational accidents involving construction scaffolding, which occurred between 2010 and 2017 in selected Polish voivodeships, were analyzed. Afterwards, the matrix of accident causes, for which the calculations were made, was created. Five subsets of accidents were obtained and the accidents were classified to a subset with similar causes.

A review of mechanical models of road pavements in the form of a proposal of classification of these models is presented. It is assumed an autonomy of the following elements of pavement model: the models of structural layers, the subgrade model, the interlayer bonding models, including bonding of pavement structure with its subgrade, the models of external impacts on pavement layers, including load of heavy traffic, the models of pavement environment impacts on structural layers’ borders (lateral) and subgrade borders (including the lower one) – according to the selected criteria such as structural criterion, material criterion (physical criterion), dimension criterion and model scope (purpose) criterion − in the frame of assumptions of the classical Newtonian deterministic mechanics. The presented attempt to classify mechanical models of road pavements supports to orientate the roadmen community within a scope of the mechanistic modelling of these structures.

This research addresses an inventory classification problem in a company that manufactures plastic pallets. Classification of the inventory is difficult because it is subject to two restrictions: the number of changeovers and the size of inventory storage. A mathematical model is first proposed to maximize the fill rate by classifying all product items into four groups. Due to all items can be classified based on the monthly demand, in descending order. The present study then proposed a procedure to find the classification that is most efficient. According to the experimental results, the maximum fill rate in the current situation is 89.85%. The proposed methodology also tested different production batches and levels of demand. The proposed methodology was found to be appropriate for practical application.

Classification of water masses in the area investigated during the 1981 FIBEX Expedition and two winter expeditions at the "H. Arctowski" Station using the method of Empirical Orthogonal Functions (EOF) is presented. Four basic water masses (warm and cold Bellinghausen Sea surface waters, surface Weddell Sea waters, Circumpolar Warm Deep Water (CWDW) and the transitional zone) were observed in the area and a significant dependence of water masses distribution ón depth was found. A strong winter increase in the Weddell Sea waters influence was recorded.

An analysis of low-level feature space for emotion recognition from the speech is presented. The main goal was to determine how the statistical properties computed from contours of low-level features influence the emotion recognition from speech signals. We have conducted several experiments to reduce and tune our initial feature set and to configure the classification stage. In the process of analysis of the audio feature space, we have employed the univariate feature selection using the chi-squared test. Then, in the first stage of classification, a default set of parameters was selected for every classifier. For the classifier that obtained the best results with the default settings, the hyperparameter tuning using cross-validation was exploited. In the result, we compared the classification results for two different languages to find out the difference between emotional states expressed in spoken sentences. The results show that from an initial feature set containing 3198 attributes we have obtained the dimensionality reduction about 80% using feature selection algorithm. The most dominant attributes selected at this stage based on the mel and bark frequency scales filterbanks with its variability described mainly by variance, median absolute deviation and standard and average deviations. Finally, the classification accuracy using tuned SVM classifier was equal to 72.5% and 88.27% for emotional spoken sentences in Polish and German languages, respectively.

Classification is one of the main areas of pattern recognition research, and within it, Support Vector Machine (SVM) is one of the most popular methods outside of the field of deep learning – and a de facto reference for many Machine Learning approaches. Its performance is determined by parameter selection, which is usually achieved by a time-consuming grid search cross-validation procedure (GSCV). That method, however, relies on the availability and quality of labelled examples and thus, when those are limited, can be hindered. To address this problem, several unsupervised heuristics exist that utilise the characteristics of the dataset to select parameters, rather than relying on class label information. While an order of magnitude faster, they are scarcely used under the assumption that their results are significantly worse than those of grid search. To challenge that assumption, we have surveyed several heuristics for SVM parameter selection and tested them against GSCV on over 30 standard classification datasets. The results demonstrate their high accuracy, with performance in terms of statistical significance comparable to GSCV, opening up an avenue for reliable label-free model defaults in resource-constrained settings e.g., edge devices or rapid prototyping.

Business Process Modelling Notation (BPMN) is a visual specification language without well-defined concepts for equivalences. This necessitates the establishment of fundamental notions that underpin the equivalences of BPMN processes. The main body of the paper is centered around the principle of substitutibility in which different types of equivalences of BPMN processes are formally described. Additionally, these results provide a basis for defining the behavioural equivalence of BPMN models. Our research investigation contributes to the field of business process management by developing a tight con-nection between BPMN and its associated equivalence notions.

A purpose of the present study is an evaluation of various models of classification of the South branch of the Cushitic languages. The South Cushitic languages are studied in their narrow sense here, i.e. without Dahalo and Ma’a, although their probable cognates are registered.

In general, currently employed vehicle classification algorithms based on the magnetic signature can distinguish among only a few vehicle classes. The work presents a new approach to this problem. A set of characteristic parameters measurable from the magnetic signature and limits of their uncertainty intervals are determined independently for each predefined class. The source of information on the vehicle parameters is its magnetic signature measured in a system that enables independent measurement of two signals, i.e. changes in the active and reactive component of the inductive loop impedance caused by a passing vehicle. These innovations result in high selective classification system, which utilizes over a dozen vehicle classes. The evaluation of the proposed approach was carried out for good vehicles consisting of 2-axle tractor and a 3-axle semi-trailer.

The aim of this study was to investigate the possible use of geoinformatics tools and generally available geodata for mapping land cover/use on the reclaimed areas. The choice of subject was dictated by the growing number of such areas and the related problem of their restoration. Modern technology, including GIS, photogrammetry and remote sensing are relevant in assessing the reclamation effects and monitoring of changes taking place on such sites. The LULC classes mapping, supported with thorough knowledge of the operator, is useful tool for the proper reclamation process evaluation. The study was performed for two post-mine sites: reclaimed external spoil heap of the sulfur mine Machów and areas after exploitation of sulfur mine Jeziórko, which are located in the Tarnobrzeski district. The research materials consisted of aerial orthophotos, which were the basis of on-screen vectorization; LANDSAT satellite images, which were used in the pixel and object based classification; and the CORINE Land Cover database as a general reference to the global maps of land cover and land use.

The systematic position of Sorbus population occurring in the Pieniny Mts. is controversial. To verify its taxonomic status we studied the ITS sequence of closely related species of the S. aria group: Sorbus sp. from the Pieniny Mts., S. aria from the Tatra Mts., S. graeca from the Balkans, and other well-distinguished native Polish Sorbus species (S. aria, S. aucuparia, S. intermedia and S. torminalis). As a reference we examined Sorbus populations closest to the Pieniny Mts. where S. graeca was reported to occur, in Slovakia. The results indicate that the Sorbus plants found in the Pieniny Mts. differ genetically from those in the Tatra Mts. but are identical to those collected from the Vihorlat Mts. in Slovakia and are closely related to S. graeca from the Balkans

This study presents the results of investigations, carried out on the Czamiawka River from December 2003 to June 2004. The results indicate the changes of physicochemical parameters of water quality. High concentration of ammonium nitrogen, COD and orthophosphates are probably caused by discharge of municipal waste-water. A drop of ammonium nitrogen, nitrite nitrogen, nitrate nitrogen concentration along the river course is probably caused by inflow of water without these components. High salinity and very high concentration of suspended solids below the "Makoszowy" coal-mine is caused by discharge of coalmine water and carbon dust from coal washer. All of the discussed parameters of water quality (except for pH-index and nitrate nitrogen) are beyond official classification. In comparison to previous analyses a slight improvement of water quality can be observed, especially in the top length. In the estuary water quality deteriorates. Although the Czarniawka River is small, it is one of the most important Kłodnica River contamination sources. Improvement of the existing situation will be possible only if firm waste-water management action will be taken, especially in the "Makoszowy" coal-mine area.

This study includes the results of investigations carried out on the Bytomka River which were made from September 2003 to January 2004. The results emphasized the changes of physico-chemical parameters of water quality. Low concentration of dissolved oxygen and high concentration of ammonium nitrogen and COD are probably caused by the discharge of municipal waste-water. High salinity is caused by coal-mines water from the river basin area. All of the discussed parameters of water quality (except for pH-index and nitrate nitrogen) are beyond official classification. Although in the river basin area there are currently activities which protect the environment, no changes of water quality have been observed except for the salinity which is growing up all the time. Improvement of the existing situation will be possible only if firm waste-water managements action is taken.

This study aims at developing a machine learning based classification and regression-based models for slope stability analysis. 1140 different cases have been analysed using the Morgenstern price method in GeoSlope for non-homogeneous cohesive slopes as input for classification and regression-based models. Slope failures presents a serious challenge across many countries of the world. Understanding the various factors responsible for slope failure is very crucial in mitigating this problem. Therefore, different parameters which may be responsible for failure of slope are considered in this study. 9 different parameters (cohesion, specific gravity, slope angle, thickness of layers, internal angle of friction, saturation condition, wind and rain, blasting conditions and cloud burst conditions) have been identified for the purpose of this study including internal, external and factors representing the geometry of the slope has been included. Four different classification algorithms namely Random Forest, logistic regression, Support Vector Machine (SVM), and K Nearest Neighbor (KNN) has been modelled and their performances have been evaluated on several performance metrics. A similar comparison based on performance indices has been made among three different regression models Decision tree, random forest, and XGBoost regression.

Synovitis is the inflammation of a synovial membrane surrounding a joint. Its assessment is an important step in the diagnosis and treatment of rheumatoid arthritis. Joint detection is the first stage of an automated method of assessment of a degree of synovitis, from an Ultrasound (USG) image of a finger joint and its surrounding area. A joint detector consists of three parts: image preprocessing, feature extraction, and classification. Each part contains adjustable parameters that must be set experimentally to ensure the proper operation of the detector. Both the structure of a joint detector and a procedure for finding a near-optimal configuration of the adjustable parameters are described. The optimization process is based on two evaluation measures: Area Under the Receiver Operating Characteristic Curve (AUC) and False Positive Count (FPC). The optimization process decreases the number of pictures with multiple detections, which was the main point of works presented in this paper. This was achieved by increasing the number of components of the homogeneous mixed-SURF descriptor which has the greatest influence on the final result. Non-SURF descriptors achieve poorer classification results. Our research led to the creation of a better joint detector which could positively influence the final results of inflammation level classification.

Knowledge about future traffic in backbone optical networks may greatly improve a range of tasks that Communications Service Providers (CSPs) have to face. This work proposes a procedure for long-term traffic forecasting in optical networks. We formulate a long-terT traffic forecasting problem as an ordinal classification task. Due to the optical networks’ (and other network technologies’) characteristics, traffic forecasting has been realized by predicting future traffic levels rather than the exact traffic volume. We examine different machine learning (ML) algorithms and compare them with time series algorithms methods. To evaluate the developed ML models, we use a quality metric, which considers the network resource usage. Datasets used during research are based on real traffic patterns presented by Internet Exchange Point in Seattle. Our study shows that ML algorithms employed for long-term traffic forecasting problem obtain high values of quality metrics. Additionally, the final choice of the ML algorithm for the forecasting task should depend on CSPs expectations.

This paper contains the results of phytosociological studies carried out on the model fragment of Spitsbergen tundra at Bellsund. In the area of 4800 m2 19 plant communities have been distinguished through association analysis and these communities, in turn, have been compared according to cluster analysis. Also, ecological groups of species have been distinguished.