Ograniczanie wyników
Czasopisma help
Autorzy help
Lata help
Preferencje help
Widoczny [Schowaj] Abstrakt
Liczba wyników

Znaleziono wyników: 158

Liczba wyników na stronie
first rewind previous Strona / 8 next fast forward last
Wyniki wyszukiwania
Wyszukiwano:
w słowach kluczowych:  clustering
help Sortuj według:

help Ogranicz wyniki do:
first rewind previous Strona / 8 next fast forward last
EN
Wireless Sensor Network finds its extensive use in healthcare applications for the transfer of time-critical data through wireless connectivity. The primary cause of network failure is the transfer of time-critical multimedia data. The article presents a new dfferentiated service modelsupported (DSM) cluster-based routing in wireless sensor networks (WSNs) that overcomes the above issue. DSM prioritizes the transfer of dfferent flow types based on packet type and packet size. The employment of computational offlading minimizes delay for critical and small-sized data packets and by carrying out data reduction of large-sized packets at proxy server. It outperforms the existing protocols in terms of energy efficiency, throughput, and reliability by prioritizing the transfer of time-critical health application data.
EN
Wireless sensor networks (WSNs) have experienced enormous growth, both in terms of the technology used and their practical applications. In order to understand the features of WSNs that make the solution suitable for a specific purpose, one needs to be aware of the theoretical concepts behind and technological aspects of WSNs. In this paper, the significance of WSNs is illustrated, with a particular emphasis placed on their demands and on understanding researchrelated problems. A review of the literature available is presented as well. Detailed discussions concerning sensor node architecture, different types of sensors used and their relevance for various types of WSNs is presented, highlighting the need to achieve application-specific requirements without degrading service quality. Multipath and cluster-based routing protocols are compared in order to analyze QoS requirements they are capable of satisfying, and their suitability for different application areas is reviewed. This survey highlights the performance of different routing protocols, therefore providing guidelines enabling each of the routing techniques to be used, in an efficient manner, with factors such as specific network structure, protocol operation and routing path construction taken into consideration in order to achieve better performance.
EN
The paper is focused on application of the clustering algorithm and Decision Tress classifier (DTs) as a semi-supervised method for the task of cognitive workload level classification. The analyzed data were collected during examination of Digit Symbol Substitution Test (DSST) with use of eye-tracker device. 26 participants took part in examination as vol-unteers. There were conducted three parts of DSST test with different levels of difficulty. As a results three versions were obtained of data: low, middle and high level of cognitive workload. The case study covered clustering of collected data by using k-means algorithm to detect three clusters or more. The obtained clusters were evaluated by three internal indices to measure the quality of clustering. The David-Boudin index detected the best results in case of four clusters. Based on this information it is possible to formulate the hypothesis of the existence of four clusters. The obtained clus-ters were adopted as classes in supervised learning and have been subjected to classification. The DTs was applied in classification. There were obtained the 0.85 mean accuracy for three-class classification and 0.73 mean accuracy for four-class classification.
PL
Celem artykułu było zastosowanie klasteryzacji wraz z klasyfikatorem Drzew Decyzyjnych jako częściowo nadzoro-wanej metody klasyfikacji poziomu obciążenia poznawczego. Dane przeznaczone do analizy zostały zebrane podczas badania DSST (z ang. Digit Symbol Substitution Test) z użyciem urządzenia eye-tracker. 26 wolontariuszów wzięło udział w badaniu. Zostały przeprowadzone trzy części testu DSST o różnych poziomach trudności. W wyniku tego, otrzymano trzy wersje danych: z niskim, średnim i wysokim poziomem obciążenia poznawczego. Do analizy danych został użyty algorytm klasteryzacji k-means do wyznaczenia trzech lub większej liczby klastrów. Uzyskane klastry zostały poddane ocenie przy użyciu trzech wewnętrznych indeksów w celu zmierzenia jakości klasteryzacji. Indeks David-Boudin’a wykazał najlepsze rezultaty w przypadku istnienia czterech klastrów. Na podstawie tej informacji można sformułować hipotezę, iż dane są podzielone na 4 klastry, co oznaczałoby istnienie dodatkowego poziomu poznawczego. Uzyskane klastry zostały zaadoptowane jako klasy w uczeniu pod nadzorem. Do klasyfikacji danych został użyty klasyfikator Drzew Decyzyjnych . Otrzymano średnią dokładność równą 0.85 w przypadku 3-klasowej klasyfikacji oraz 0.73 średnią dokładność dla 4-klasowej klasyfikacji.
EN
Automated essay evaluation is a widely used practical solution for replacing time-consuming manual grading of student essays. Automated systems are used in combination with human graders in different high-stake assessments, where grading models are learned on essays datasets scored by different graders. Despite the definition of the standardized grading rules, human graders can unintentionally introduce subjective bias into scores. Consequently, a grading model has to learn from data that represents a noisy relationship between essay attributes and its grade. We propose an approach for partitioning a set of essays into subsets that represent similar graders, which uses an explanation methodology and clustering. The results confirm our assumption that learning from the ensemble of separated models can significantly improve the average prediction accuracy on artificial and real-world datasets.
EN
The paper presents an ensemble classification method based on clustering, along with its implementation in the Python programming language. An illustrative example showing the method behavior is provided, and the results of a computational experiment performed on real life data sets are reported.
EN
The problem of evaluation of decisions is considered, which evaluation consists in selecting from the set of possible decisions those that meet the decision-maker's preferences. The added value of solving this problem lies in the reduction of the number of decisions one can choose. Evaluation of decisions is based on their complete characteristics, rather than on a pre-defined quality indicator. The basis for the quality assessment are given pattern examples of decisions made. These are decisions that the decision maker has found to be exemplary or acceptable. They are used as defining his preferences. The methods proposed in this article concern the ordering and clustering of decisions based on their characteristics. The set of decisions selected by an algorithm is interpreted as recommended for the decision maker. Presented solutions can find a variety of applications, for example in investment planning, routing, diagnostics or searching through multimedia databases.
PL
Rozpatrywany jest problem ewaluacji decyzji polegający na wytypowaniu spośród możliwych decyzji tych, które spełniają preferencje decydenta. Użyteczność rozwiązania problemu polega na zredukowaniu liczby możliwych do wyboru decyzji. Ewaluacja decyzji bazuje na ich kompletnych charakterystykach, a nie na wcześniej zdefiniowanym wskaźniku jakości. Podstawą oceny jakości są wzorcowe przykłady decyzji. Są to decyzje, które decydent uznał za doskonałe lub akceptowalne. Wskazane przez decydenta przykłady są wykorzystywane jako określające jego preferencje. Proponowane w artykule metody dotyczą porządkowania i grupowania decyzji na podstawie ich charakterystyk. Wytypowany zbiór decyzji jest interpretowany jako rekomendowany dla decydenta. Przedstawione rozwiązania mogą znaleźć różnorakie zastosowania, np. w planowaniu inwestycji, trasowaniu, diagnostyce czy przeszukiwaniu multimedialnych baz danych.
EN
The paper presents the proposed protocol a hybrid approach is applied for clustering of sensor networks combining BBO and K-means algorithm. The performance of the protocol is compared with SEP, IHCR and ERP in terms of stability period, network life time, residual energy and throughput. The simulation results show that the proposed protocol named as KBBO has improved the performance of these parameters significantly.
PL
W pracy przedstawiono protokół, w którym stosuje się podejście hybrydowe do grupowania sieci czujników łączących algorytm BBO i K-średnich. Jego wydajność jest porównywana z SEP, IHCR i ERP pod względem okresu stabilności, żywotności sieci, energii resztkowej I przepustowości. Wyniki symulacji pokazują, że prezentowany protokół nazwany KBBO znacznie poprawił wydajność tych parametrów.
8
EN
Clustering algorithms are usually based on an initial estimate of cores, have performance dependent on the number of clusters and dimension of the data, and are performed offline. Thus, by categorizing a highly coupled sensor network as an industrial plant, it is necessary that all these characteristics are achieved. The article presents an improvement of the TEDA-Cloud, based on the Typicity and Eccentricity Data Analitics (TEDA).Inthisway,theproposed(TEDA-Cloudmodified),methodreducestheamountofstoreddataformergingcoresandspeedsuptheclassification of the presented data.
PL
Algorytmy klastrowania są zwykle oparte na wstępnym oszacowaniu rdzeni, mają wydajność zależną od liczby klastrów i wymiarów danych i są wykonywane w trybie offline. Zatem, poprzez kategoryzowanie wysoce sprzężonej sieci czujników jako instalacji przemysłowej, konieczne jest, aby wszystkie te cechy zostały osiągnięte. W artykule przedstawiono ulepszenie chmury TEDA opartej na analizie typowości i ekscentryczności danych (TEDA). W ten sposób proponowana (zmodyfikowana TEDA-Cloud) metoda zmniejsza ilość przechowywanych danych do łączenia rdzeni i przyspiesza klasyfikację prezentowanych danych.
EN
Advanced and accurate modelling of a Flapping Wing Micro Air Vehicle (FW MAV) and its control is one of the recent research topics related to the field of autonomous MAVs. Some desiring features of the FW MAV are quick flight, vertical take-off and landing, hovering, and fast turn, and enhanced manoeuvrability contrasted with similar-sized fixed and rotary wing MAVs. Inspired by the FW MAV’s advanced features, a four-wing Natureinspired (NI) FW MAV is modelled and controlled in this work. The Fuzzy C-Means (FCM) clustering algorithm is utilized to construct the data-driven NIFW MAV model. Being model free, it does not depend on the system dynamics and can incorporate various uncertainties like sensor error, wind gust etc. Furthermore, a Takagi-Sugeno (T-S) fuzzy structure based adaptive fuzzy controller is proposed. The proposed adaptive controller can tune its antecedent and consequent parameters using FCM clustering technique. This controller is employed to control the altitude of the NIFW MAV, and compared with a standalone Proportional Integral Derivative (PID) controller, and a Sliding Mode Control (SMC) theory based advanced controller. Parameter adaptation of the proposed controller helps to outperform it static PID counterpart. Performance of our controller is also comparable with its advanced and complex counterpart namely SMC-Fuzzy controller.
EN
The paper presents a short-time analysis of the vibration signals for the diagnosis of Diesel engine of combustion locomotive by recognition of different engine states using the clustering technique. The main aim of the researches was to distinguish between different engine states represent different wear extends. The proposed method of vibration signal analysis consists on sliding a time window along signal in time and observing the changes of some given statistical parameters. The set of this parameter values creates a multidimensional parameter space where the time evolution can be observed. For recognition and detection of different engine system states some clustering techniques in the parameter space were performed. The results show the possibility of distinguishing different cluster centers within the parameter space which can be assigning to different engine states represented the states before and after a general repair.
EN
The properties of hypoeutectic Al–Si alloy (silumin) with the addition of elements such as Cr, Mo, V and W are described. Changes in silumin microstructure under the impact of these elements result in a change of the mechanical properties. The research includes presentation of procedure for the acquisition of knowledge about these changes directly from experimental results using mixed data mining techniques. The procedure for analyzing small sets of experimental data for multistage, multivariate and multivariable models has been developed. Its use can greatly simplify such research in the future. An interesting achievement is the development of a voting procedure based on the results of classification trees and cluster analysis.
EN
This review paper presents a shortcoming associated to data mining algorithm(s) classification, clustering, association and regression which are highly used as a tool in different research communities. Data mining researches has successfully handling large amounts of dataset to solve the problems. An increase in data sizes was brought a bottleneck on algorithms to retrieve hidden knowledge from a large volume of datasets. On the other hand, data mining algorithm(s) has been unable to analysis the same rate of growth. Data mining algorithm(s) must be efficient and visual architecture in order to effectively extract information from huge amounts of data in many data repositories or in dynamic data streams. Data visualization researchers believe in the importance of giving users an overview and insight into the data distributions. The combination of the graphical interface is permit to navigate through the complexity of statistical and data mining techniques to create powerful models. Therefore, there is an increasing need to understand the bottlenecks associated with the data mining algorithms in modern architectures and research community. This review paper basically to guide and help the researchers specifically to identify the shortcoming of data mining techniques with domain area in solving a certain problems they will explore. It also shows the research areas particularly a multimedia (where data can be sequential, audio signal, video signal, spatio-temporal, temporal, time series etc) in which data mining algorithms not yet used.
EN
Unconventional oil and gas reservoirs from the lower Palaeozoic basin at the western slope of the East European Craton were taken into account in this study. The aim was to supply and improve standard well logs interpretation based on machine learning methods, especially ANNs. ANNs were used on standard well logging data, e.g. P-wave velocity, density, resistivity, neutron porosity, radioactivity and photoelectric factor. During the calculations, information about lithology or stratigraphy was not taken into account. We apply different methods of classification: cluster analysis, support vector machine and artificial neural network—Kohonen algorithm. We compare the results and analyse obtained electrofacies. Machine learning method–support vector machine SVM was used for classification. For the same data set, SVM algorithm application results were compared to the results of the Kohonen algorithm. The results were very similar. We obtained very good agreement of results. Kohonen algorithm (ANN) was used for pattern recognition and identification of electrofacies. Kohonen algorithm was also used for geological interpretation of well logs data. As a result of Kohonen algorithm application, groups corresponding to the gas-bearing intervals were found. Analysis showed diversification between gas-bearing formations and surrounding beds. It is also shown that internal diversification in gas-saturated beds is present. It is concluded that ANN appeared to be a useful and quick tool for preliminary classification of members and gas-saturated identification.
14
Content available remote A concept of detection method for botnets based on social networks
EN
There are a lot of botnets implementations which are using different kind of communication protocols such as P2P, HTTP, IRC. There are also a lot of methods of their detection which are in most cases useless against botnets that are using novel communication protocols. In nowadays, one can observe increasing number of Internet threats that are using new kind of communication methods for receiving and sending commands between infected host and botmaster. The aim of this paper is to present a concept of detection method for botnets that are using social networks for communication with Command & Control.
PL
Obecnie istnieje wiele implementacji botnetów różniących się przede wszystkim wykorzystywanym protokołem komunikacji, np.: P2P, HTTP, IRC. W związku z powyższym powstały liczne metody ich wykrywania. Niestety znaczna ich część jest nieskuteczna wobec zagrożeń wykorzystujących nowatorskie metody komunikacji. Celem niniejszego artykułu jest zaprezentowanie metody pozwalającej na wykrycie botnetów, które wykorzystują sieci społecznościowe na zarządzania farmami zainfekowanych komputerów.
PL
W referacie przedstawiono wyniki badań implementacyjnych i symulacyjnych nad technikami detekcji zajętości zasobów widmowych, charakteryzującymi się podwyższoną wiarygodnością i/lub oszczędnością energetyczną. W tym celu dokonano implementacji metod detekcji na urządzeniach i oceniono wpływ praktycznych urządzeń na wyniki detekcji. Ponadto, oszacowano zyski płynące z kooperacyjnej detekcji sygnałów także w środowisku wysokiej korelacji. Zidentyfikowano możliwe kierunki energooszczędnej kooperacyjnej detekcji sygnałów i wyznaczono zależności między kierunkami optymalizacji. Ostatecznie, zaproponowano energooszczędne techniki detekcji w szczególności oparte na wyborze i grupowaniu węzłów.
EN
In the paper the results of implementation- and simulation-based spectrum sensing methods with enhanced reliability and/or energy efficiency have been presented. To this aim, selected spectrum sensing techniques have been implemented and the impact of hardware has been assessed. Moreover, the gains which support the cooperation scheme have been assessed including correlation-based scheme. Possible directions in energy-efficient cooperative spectrum sensing have been identified and the dependencies between directions of optimization have been drawn. Finally, the energy-efficient techniques have been developed with the use of node selection and clustering.
EN
The paper presents the results of research on the impact of currency regime type on features of the spread of financial crises. The focus is on constructing a neural network to identify groups of countries exhibiting similar behaviour in the dynamics of the index of flexibility in the effective exchange rate, exchange market pressure and external public debt markets in times of sudden changes in the environment. The alpha-criterion for optimality constructed in this way is based on the use of a concordance coefficient. The result of modelling is a self-organization map with a hidden layer consisting of six clusters. This cluster structure allows us to analyse the relationship between the type of currency regime and the consequences of the global crisis in 2007–2009 for the domestic financial markets of the investigated countries. It is found that the result of the division is significantly influenced by the proximity of administrative boundaries and historically predetermined close trade and economic channels of interaction between economies. The results obtained can be used to formulate directions in the currency policies of developing countries, including Ukraine.
EN
This paper proposes a fuzzy Manhattan distance-based similarity for gang formation of resources (FMDSGR) method with priority task scheduling in cloud computing. The proposed work decides which processor is to execute the current task in order to achieve efficient resource utilization and effective task scheduling. FMDSGR groups the resources into gangs which rely upon the similarity of resource characteristics in order to use the resources effectively. Then, the tasks are scheduled based on the priority in the gang of processors using gang-based priority scheduling (GPS). This reduces mainly the cost of deciding which processor is to execute the current task. Performance has been evaluated in terms of makespan, scheduling length ratio, speedup, efficiency and load balancing. CloudSim simulator is the toolkit used for simulation and for demonstrating experimental results in cloud computing environments.
EN
This article presents the possibilities for using cluster analysis in the assignment of machine tools in automated manufacturing systems. Based on the similarity of manufacturing processes in the system, cutting tools have been grouped. The objective was to obtain groups of similar objects, which could potentially ensure the reduction of the frequency and time of setups, optimizing the maintenance of tool resources and improving the efficiency and quality of production. With the application of similarity coefficients and hierarchical clustering algorithms, tool sets were formed with their composition specified. The assumed key factor was the limited tool magazine capacity for the machine tool. Therefore, it was necessary to separate the group with the largest multiplicity, not exceeding the assumed tool magazine capacity, from each group. The final part of the study includes an evaluation of the obtained solutions with selected measures used.
PL
W niniejszym artykule przedstawiono możliwości zastosowania analizy skupień w przydziale narzędzi do obrabiarek w zautomatyzowanych systemach wytwarzania. Bazując na podobieństwie używanych w systemie procesów wytwórczych grupowaniu, poddano narzędzia obróbkowe. Celem było uzyskanie grup obiektów podobnych, które potencjalnie zapewnić mogły zmniejszenie liczby i czasu przezbrojeń, lepsze wykorzystanie zasobu narzędziowego oraz poprawę efektywności i jakości produkcji. Z wykorzystaniem współczynników podobieństwa i hierarchicznych algorytmów grupowania stworzono zestawy narzędziowe i określono ich skład. Jako czynnik kluczowy przyjęto ograniczoną pojemność magazynu narzędziowego obrabiarki. Koniecznym stało się zatem wyodrębnienie z każdej możliwej liczby grup grupy o największej liczności, która nie przekraczała założonej pojemności magazynu narzędziowego. W ostatniej części opracowania przeprowadzono ocenę uzyskanych rozwiązań z wykorzystaniem wybranych miar.
EN
Analysis of electrocardiogram and heart rate provides useful information about health condition of a patient. The North Sea Bicycle Race is an annual cycling competition in Norway. Examination of ECG recordings collected from participants of this race may allow defining and evaluating the relationship between physical endurance exercises and heart electrophysiology. Parameters reflecting potentially alarming deviations are to be identified in this study. This paper presents results of a time-domain analysis of ECG data collected in 2014, implementing K-Means clustering. A double stage analysis strategy, aimed at producing hierarchical clusters, is proposed. The first phase allows rough separation of data. Second stage is applied to reveal internal structure of the majority clusters. In both steps, discrepancies driving the separation could stem from three sources. Firstly, they could be signs of abnormalities in electrical activity of the heart. Secondly, they may allow discriminating between natural groups of participants – according to sex, age, physical fitness. Finally, some deviations could result from faults in data extraction, therefore serving in evaluation of the parameters. The clusters were defined predominantly by combinations of features: heartbeat signals correlation, P-wave shape, and RR intervals; none of the features alone was discriminative for all the clusters.
EN
The article presents an assessment of the maximum Hello period value required for maintaining the assumed network awareness in the context of available transmitter power, number of nodes and their velocity. The Lowest ID algorithm was used for defining the node functions. This problem is especially important for low density and high mobility networks. Assuming that 1% of network nodes have the wrong knowledge about their states, the length of the Hello period in a typical system behavior cannot be higher than 0.1 to 0.5 seconds.
PL
W artykule przedstawiono szacunek czasu powtarzania wiadomości Hello, który zapewnia założony poziom świadomości sieci. W ocenie wzięto pod uwagę zmienność mocy nadawania węzłów, ich prędkość przemieszczania się oraz liczbę. Badania przeprowadzono przy wykorzystaniu algorytmu klasteryzacji Lowest ID. Prezentowany problem jest szczególnie ważny dla sieci mobilnych o małej gęstości. Ocenia się, że dla typowych zastosowań czas powtarzania wiadomości Hello nie powinien być dłuższy niż 0,1 do 0,5 sekundy.
first rewind previous Strona / 8 next fast forward last
JavaScript jest wyłączony w Twojej przeglądarce internetowej. Włącz go, a następnie odśwież stronę, aby móc w pełni z niej korzystać.