Wyniki wyszukiwania - BazTech

1

An empirical study of a simple incremental classifier based on vector quantization and adaptive resonance theory

Czmil Sylwester, Kluska Jacek, Czmil Anna

International Journal of Applied Mathematics and Computer Science

|

2024

|

Vol. 34, no. 1

149--165

EN

When constructing a new data classification algorithm, relevant quality indices such as classification accuracy (ACC) or the area under the receiver operating characteristic curve (AUC) should be investigated. End-users of these algorithms are interested in high values of the metrics as well as the proposed algorithm’s understandability and transparency. In this paper, a simple evolving vector quantization (SEVQ) algorithm is proposed, which is a novel supervised incremental learning classifier. Algorithms from the family of adaptive resonance theory and learning vector quantization inspired this method. Classifier performance was tested on 36 data sets and compared with 10 traditional and 15 incremental algorithms. SEVQ scored very well, especially among incremental algorithms, and it was found to be the best incremental classifier if the quality criterion is the AUC. The Scott–Knott analysis showed that SEVQ is comparable in performance to traditional algorithms and the leading group of incremental algorithms. The Wilcoxon rank test confirmed the reliability of the obtained results. This article shows that it is possible to obtain outstanding classification quality metrics while keeping the conceptual and computational simplicity of the classification algorithm.

2

Selection of an algorithm for classifying data quoted on the Day Ahead Market of TGE S.A. in MATLAB and Simulink using Deep Learning Toolbox

Tchórzewski Jerzy, Mielcarz Tomasz

Studia Informatica : systems and information technology

|

2023

|

Vol. 1(28)

83--108

EN

The article contains an analysis leading to the selection of an algorithm for classifying data listed on the Day-Ahead Market of TGE S.A. in MATLAB and Simulink using Deep Learning Toolbox. In this regard, an introduction to deep learning methods, classification methods, and classification algorithms is provided first. Particular attention was paid to the essence of three important deep learning methods in the classification, i.e. the methods called: Stochastic Gradient Descent Momentum, Root Mean Square Prop and Adaptive Moment Estimation. Then, three architectures of artificial neural networks used in deep learning were characterized, i.e.: Deep Belief Network, Convolutional Neural Network and Recurrent Neural Network. Attention was paid to the selection parameters of algorithms for learning deep artificial neural networks that can be used in classification, such as: accuracy, information losses and learning time. Practical aspects of research experiments were also shown, including selected results of research conducted on volume and fixing 1 data quoted on the TGE S.A. Day-Ahead Market. After analyzing the obtained test results for the hourly system, it was noted that the least suitable algorithm for classification purposes was the Stochastic Gradient Descent Momentum algorithm, which in each case had worse results than the other two algorithms, i.e. the Adaptive Moment Estimation algorithm and the Root Mean algorithm Square Prop. However, the best algorithm turned out to be the Adaptive Moment Estimation algorithm, which obtained the highest accuracy, which was at a level comparable to the Root Mean Square Prop algorithm, with the latter algorithm having larger losses.

3

Enhanced algorithm for energy optimization and improvised synchronization in knee exoskeleton system

Arunamithra J., Saravanan R., Venkatesh Babu S.

Archives of Materials Science and Engineering

|

2022

|

Vol. 117, nr 2

79--85

EN

Purpose: The purpose of the study is to develop an augmented algorithm with optimised energy and improvised synchronisation to assist the knee exoskeleton design. This enhanced algorithm is used to estimate the accurate left and right movement signals from the brain and accordingly moves the lower-limb exoskeleton with the help of motors. Design/methodology/approach: An optimised deep learning algorithm is developed to differentiate the right and left leg movements from the acquired brain signals. The obtained test signals are then compared with the signals obtained from the conventional algorithm to find the accuracy of the algorithm. Findings: The obtained average accuracy rate of about 63% illustrates the improvised differentiation in identifying the right and left leg movement. Research limitations/implications: The future work involves the comparative study of the proposed algorithm with other classification technologies to extract more reliable results. A comparative analysis of the replaceable and rechargeable battery will be done in the future study to exhibit the effectiveness of the proposed model. Originality/value: This study involves the extended study of five frequency regions namely alpha, beta, gamma, delta and theta, to handle the real-time EEG signal processing exoskeleton, model.

4

Data classification to improve the algorithm for assessing an external patchwork of land ownership

Oleniacz Grzegorz, Noga Karol, Moskal Adrian

Geomatics, Landmanagement and Landscape

|

2021

|

no. 4

75--86

EN

Research conducted for many years in Poland and around the world has demonstrated that defects in the spatial structure of agricultural land resulting from the common phenomenon of land fragmentation constitute one of important factors that contribute to the lack of rational land management. Reconstruction of the spatial structure of rural areas is essential for sustainable development of these areas. The process of land consolidation and exchange is a tool that can arrange space and lead to the desired structural changes, however, it has to be performed systematically and primarily in those areas where it is most urgently needed. With limited budget resources, it is reasonable to select objects for land consolidation in such a way as to obtain the best possible effect. In case of a selected group of neighbouring villages, a joint land consolidation procedure will enable better consolidation effects, inter alia by eliminating the external patchwork of land ownership. This article describes the modification and improvement in the methodology of designating areas with concentration of the external patchwork of land ownership. It is based on Czekanowski’s diagram and was originally presented in 2017. The applied classification of numerical data enables the elimination of undesirable numerical effects of calculations and simplifies the interpretation of the final results.

5

Basic quantum circuits for classification and approximation tasks

Wiśniewska Joanna, Sawerwain Marek, Obuchowicz Andrzej

International Journal of Applied Mathematics and Computer Science

|

2020

|

Vol. 30, no. 4

733--744

EN

We discuss a quantum circuit construction designed for classification. The circuit is built of regularly placed elementary quantum gates, which implies the simplicity of the presented solution. The realization of the classification task is possible after the procedure of supervised learning which constitutes parameter optimization of Pauli gates. The process of learning can be performed by a physical quantum machine but also by simulation of quantum computation on a classical computer. The parameters of Pauli gates are selected by calculating changes in the gradient for different sets of these parameters. The proposed solution was successfully tested in binary classification and estimation of basic non-linear function values, e.g., the sine, the cosine, and the tangent. In both the cases, the circuit construction uses one or more identical unitary operations, and contains only two qubits and three quantum gates. This simplicity is a great advantage because it enables the practical implementation on quantum machines easily accessible in the nearest future.

6

Comparing continuity and compactness of choropleth map classes

Całka B.

Geodesy and Cartography

|

2018

|

Vol. 67, no. 1

21--34

EN

Population density varies sharply from place to place on the whole territory of Poland. The largest number of people per 1 km2 is 21,531, while uninhabited areas account for about 48% of the country. Such uneven, non-Gaussian distribution of the data causes some difficulty in choosing the classification method in geometric choropleth maps. A thorough evaluation of a geometric choropleth map of population data is not possible using only traditional indicators such as the Tabular Accuracy Index (TAI). That is why the aim of the article is to develop an innovative index based on distance analysis and neighbour analysis of grid cells. Two indexes have been suggested in this paper: the Spatial Distance Index (SDI) and the Spatial Contiguity Index (SCI). The paper discusses the use of five classification methods to evaluate choropleth maps of population data, like head-tail breaks, natural breaks, equal intervals, quantile, and geometrical intervals. A comprehensive assessment of such geometric choropleth maps is also done. The research was conducted for the whole territory of Poland, using data from the 2011 National Census of Population and Housing. Population data are presented in the 1km grid. The results of the analysis are shown on thematic maps. A compatibility of the choropleth maps with urban-rural typology of the OECD (Organisation for Economic Co-operation and Development) was also checked.

7

Badanie jakości predykcyjnej segmentacji rynku

Paśko Ł., Setlak G.

Studia Informatica

|

2016

|

Vol. 37, nr 1

83--97

PL

Celem pracy jest ocena wyników predykcyjnej segmentacji rynku za pomocą narzędzi wykorzystywanych do badania jakości klasyfikatorów. Omawiana predykcyjna segmentacja rynku dotyczyła wyrobów gospodarstwa domowego. Przeprowadzono ją, wykorzystując klasyfikatory CART i CHAID. W pracy przedstawiono rezultaty oceny tych klasyfikatorów oraz wynikające z tego wnioski, dotyczące jakości segmentacji rynku.

EN

The aim of the paper is to assess the results of predictive market segmentation using methods of examination of classifiers’ quality. The discussed predictive market segmentation was applied to household products. It was performed using CART and CHAID classifiers. The article contains the results of assessing the classifiers and the consequent conclusions on the quality of market segmentation.

8

Klasyfikacja przemieszczeń pionowych obiektów budowlanych z wykorzystaniem sieci neuronowej SVM

Mrówczyńska M.

Materiały Budowlane

|

2016

|

nr 11

44--45

PL

W artykule przedstawiono podstawowe zasady budowania i uczenia sieci neuronowych zwane metodą (techniką) wektorów podtrzymujących (ang. Support Vector Machine – SVM) wraz z możliwością aplikacji tego rodzaju sieci. Sieci nieliniowe SVM wykorzystano do klasyfikacji danych separowalnych liniowo w celu sformułowania modelu przemieszczeń punktów sieci pomiarowo-kontrolnej. Punkty sieci pomiarowo-kontrolnej zostały założone na obiekcie budowlanym posadowionym na gruntach ekspansywnych.

EN

The article presents the basic rules for constructing and training neural networks called the Support Vector Machine (SVM) method as well as possible applications for this kind of network. Non-linear SVM networks have been used for classifying linearly separable data with a view to formulating a model of displacements of points in a measurement-control network. The points of the measurement-control network were placed on a civil engineering object located on expansive soil.

9

Medical data preprocessing for increased selectivity of diagnosis

Walczak A., Paczkowski M.

Bio-Algorithms and Med-Systems

|

2016

|

Vol. 12, no. 1

39--43

EN

In this review, we present a framework that will enable us to obtain increased accuracy of computer diagnosis in medical patient checkups. To some extent, a new proposition for medical data analysis has been built based on medical data preprocessing. The result of such preprocessing is transformation of medical data from descriptive, semantic form into parameterized math form. A proper model for digging of hidden medical data properties is presented as well. Exploration of hidden data properties achieved by means of preprocessing creates new possibilities for medical data interpretation. Diagnosis selectivity has been increased by means of parameterized illnesses patterns in medical databases.

10

Recognition of lip prints using Fuzzy c-Means clustering

Wrobel K., Froelich W.

Journal of Medical Informatics & Technologies

|

2015

|

Vol. 24

67--73

EN

In this paper a new method for lip print recognition is proposed. The proposed approach is based on Fuzzy c-Means clustering of the characteristics features of lip prints. First, the Hough transform is applied for the recognition of the characteristic features within lip prints, then Fuzzy c-Means clustering is performed to cluster those features. The proposed algorithm applies the results of clustering to find an unknown image withing the collected repository of lip prints. Instead of comparing all pairs of individual characteristic features, the proposed algorithm uses the representatives of clusters for the comparison of images. The advantage of using the proposed method is its increased tolerance to the noise in data and thus the increased efficiency of the recognition. The effectiveness of presented method has been verified experimentally using real-world images. The results are satisfactory and suggest the possibility of using the method in forensic identification systems

11

Personal identity verification method based on lips photographs

Wrobel K., Doroz R., Porwik P., Naruniec J., Kowalski M.

Journal of Medical Informatics & Technologies

|

2015

|

Vol. 24

59--65

EN

The paper presents a personal identification method based on lips photographs. This method uses a new approach to the extraction and classification of characteristic features of the mouth from photographs. It eliminates the drawbacks that occur during the acquisition of lip print images with the use of the forensic method that requires special tools. Geometrical dimensions of the entire mouth as well as of the upper and lower lips were adopted as the features, on the basis of which the verification is performed. An ensemble classifier was used for the classification of the features obtained. The effectiveness of the classifier has been verified experimentally.

12

Towards gesture recognition in three-dimensional space

Gadomer Ł.

Advances in Computer Science Research

|

2015

|

Nr 12

5--20

EN

In this work, author describes the continuation of his researches about gesture recognition. The previous varaint of the solution was using plain data and was dependent of the performance velocity. In the described researches author made it speed and position invariant by resolving problem of too long or too short gestures – in a previous solution the user had to decide about gesture duration time before performing, now it is not necessary. He also proposed another data representations, using features computed of recorded data. Previous representation, which assumed storing relative positions between samples, was replaced by transforming each gesture to the axis origin and normalizing. He also tried to connect these two representations – plain data and features – into a single one. All of these new data representations were tested using the SVM classifier, which was judged to be the best for the given problem in the previous work. Each of them was tested using one of four popular SVM kernel functions: linear, polynomial, sigmoid and radial basis function (RBF). All achieved results are presented and compared.

PL

W niniejszym artykule autor opisał kontynuację swoich badań dotyczących rozpoznawania gestów. Ulepszył on stworzone przez siebie rozwiązanie w taki sposób, aby nagrywanie i rozpoznawanie gestów było niezależne od szybkości ich wykonywania, a co za tym idzie — ich zróżnicowanej długości. Zaproponował on także inne reprezentacje danych, za pomocą których wyrażany jest stworzony zbiór gestów. Wcześniejsze rozwiązanie, opierające się na przechowywaniu relatywnego położenia dłoni w stosunku do poprzedniej zarejestrowanej próbki (poprzedniego położenia), zastąpione zostało sprowadzeniem gestu do początku układu współrzędnych i zastąpieniem wartości relatywnych absolutnymi, a następnie ich normalizację Z tak przygotowanego zbioru gestów obliczone zostały cechy stanowiące drugą zaproponowaną reprezentację danych. Trzecia reprezentacja stanowi połączenie dwóch poprzednich: zawiera jednocześnie bezpośrednie wartości wyrażające ruch dłoni, jak i obliczone na podstawie jego cechy. Wszystkie trzy reprezentacje zostały przetestowane przy pomocy klasyfikatora, który okazał się najlepszy dla zadanego problemu podczas przeprowadzania wcześniejszych badań: SVM. Porównano, jak z zadanym problemem radzą sobie cztery popularne funkcje jądra: liniowa, wielomianowa, sigmoidalna i radialna. Otrzymane wyniki zostały przedstawione, porównane i omówione.

13

Comparative Study of Supervised Learning Methods for Malware Analysis

Kruczkowski M., Niewiadomska-Szynkiewicz E.

Journal of Telecommunications and Information Technology

|

2014

|

nr 4

24--33

EN

Malware is a software designed to disrupt or even damage computer system or do other unwanted actions. Nowadays, malware is a common threat of the World Wide Web. Anti-malware protection and intrusion detection can be significantly supported by a comprehensive and extensive analysis of data on the Web. The aim of such analysis is a classification of the collected data into two sets, i.e., normal and malicious data. In this paper the authors investigate the use of three supervised learning methods for data mining to support the malware detection. The results of applications of Support Vector Machine, Naive Bayes and k-Nearest Neighbors techniques to classification of the data taken from devices located in many units, organizations and monitoring systems serviced by CERT Poland are described. The performance of all methods is compared and discussed. The results of performed experiments show that the supervised learning algorithms method can be successfully used to computer data analysis, and can support computer emergency response teams in threats detection.

14

Diagnosing Parkinson’s disease using the classification of speech signals

Froelich W., Wróbel K., Porwik P.

Journal of Medical Informatics & Technologies

|

2014

|

Vol. 23

187--193

EN

This paper addressees the problem of an early diagnosis of Parkinson’s disease by the classification of characteristic features of person’s voice. A new, two-step classification approach is proposed. In the first step, the voice samples are classified using standard state-of-the-art classifiers. In the second step, the classified samples are assigned to patients and the final classification process based on majority criterion is performed. The advantage of using our new approach is the resulting, reliable patientoriented medical diagnose. The proposed two-step method of classification allows also to deal with the variable number of voice samples gathered for every patient. Preliminary experiments revealed quite satisfactory classification accuracy obtained during the performed leave-one-out cross validation.

15

The detection of selected marine engine malfunctions on the basis of the exhaust gas composition

Kowalski J.

Diagnostyka

|

2014

|

Vol. 15, No. 3

39--44

EN

Marine diesel engines are generators of mechanical energy, but also are generators of toxic compounds into the atmosphere. The composition of the exhaust gas may be a carrier of diagnostic information about the condition of functional systems of the engine. Results of the classification and selection of diagnostic signals for selected marine engine malfunctions are presented. The analysis was based on results of laboratory tests. Mentioned classification was able to isolate symptoms of malfunctions of marine 4-stroke diesel engine in the composition of the exhaust gas. Complementary detection signals are exhaust gas temperature behind each cylinder. The conclusion of this work is the ability to detect by this method such engine malfunctions as the throttling the air intake duct, the throttling of the exhaust gas duct, the decreasing and the increasing of fuel injection pressure on the selected cylinder, chocked or discalibrated fuel injector, the leakage of the fuel pump, changing of the fuel injection timing and exhaust and inlet valves malfunctions.

PL

Silniki tłokowe generatorami energii mechanicznej, ale również generatorami emisji związków toksycznych do atmosfery. Skład emitowanych spalin może być nośnikiem informacji diagnostycznej o stanie technicznym układów funkcjonalnych silnika. W pracy przedstawiono wyniki klasyfikacji i wyboru sygnałów diagnostycznych dla wybranych niesprawności silnika okrętowego. Analiza została oparta na wynikach badań laboratoryjnych. W wyniku przeprowadzonych działań udało się wyodrębnić symptomy niesprawności 4-suwowego silnika okrętowego w składzie emitowanych spalin. Sygnałami uzupełniającymi detekcję są temperatury gazów spalinowych za poszczególnymi cylindrami. Wnioskiem z prezentowanej pracy jest możliwość wykrycia tą metodą takich niesprawności silnika jak dławienie kanału dolotowego powietrza i wylotowego spalin, obniżenie i zwiększenie ciśnienia wtrysku paliwa do wybranego cylindra, zakoksowanie lub rozkalibrowanie wtryskiwacza, przecieki w pompie paliwowej, zmiana rozpoczęcia wtrysku paliwa oraz wypalenie gniazd zaworów dolotowych i wylotowych.

16

Zastosowanie metod eksploracji danych do segmentacji rynków

Setlak G., Paśko Ł.

Studia Informatica

|

2013

|

Vol. 34, nr 2A

311--323

PL

Celem niniejszego artykułu są przedstawienie i ocena możliwości wykorzystania metod eksploracji danych do segmentacji rynków zbytu. Przedstawiono segmentacje opisową i predykcyjną oraz przeanalizowano wyniki rozwiązywania zadań klasyfikacji i grupowania danych za pomocą sieci neuronowych Kohonena oraz drzew klasyfikacyjnych CART i CHAID. W pracy wykorzystano dane dotyczące rynków zbytu przedsiębiorstwa produkującego wyroby gospodarstwa domowego.

EN

The purpose of this paper is to present and evaluate the possibility of using data mining methods in the market segmentation process. In the paper the descriptive and predictive segmentation were presented and the results of classification and clustering data were analyzed. To carry out the analysis were used following methods: Kohonen neural networks, CART and CHAID. The analysis concerns the manufacturing company producing household products.

17

Konstrukcja systemów eksploracji danych dla obrazów rastrowych

Ładniak M., Piórkowski A., Młynarczuk M.

Studia Informatica

|

2013

|

Vol. 34, nr 2B

7--20

PL

Poniższy artykuł porusza tematykę możliwości wydobywania wiedzy zawartej w obrazach rastrowych, magazynowanych w bazie danych. Zaproponowano autorskie podejście do możliwości konstrukcji systemu wspomagającego proces eksploracji takich danych. Omówione zostały aspekty konstrukcyjne oraz implementacyjne.

EN

The following article deals with possibilities for retrieving information from raster images stored in a database. An author’s approach for the possibility of a support system construction for the process of exploration of such data is proposed. The following article describes construction and implementation aspects.

18

Analysis of medical data using dimensionality reduction techniques

Siwek K., Osowski S., Markiewicz T., Korytkowski J.

Przegląd Elektrotechniczny

|

2013

|

R. 89, nr 2a

279--281

EN

The paper presents the application of dimensionality reduction methods for representation of the multidimensional medical data representing the images of the blood cells in leukemia. Different techniques of reduction belonging to linear and nonlinear methods will be applied and their efficiency compared. Their application to the visualization of different classes as well as clusterization and classification of data will be studied and discussed in the paper.

PL

Praca przedstawia zastosowanie różnych metod redukcji wymiaru danych w reprezentacji numerycznej deskryptorów charakteryzujących klasy komórek krwiotwórczych w białaczce. Porównane zostaną różne podejścia do redukcji oparte na metodach liniowych i nieliniowych transformacji. W szczególności analizie poddane zostaną możliwości zastosowania tych metod w wizualizacji danych jak również klasteryzacji i klasyfikacji. W pracy pokazane zostaną wyniki przeprowadzonych eksperymentów dotyczących 11 klas komórek.

19

Application of fussion classify for data classification

Duraj A.

Journal of Applied Computer Science

|

2013

|

Vol. 21, nr 1

39--52

EN

In the published articles and works there are solutions regarding data fusion. However, there is not any verification as for the efficiency of classifiers in the case of many sources of data given simultaneously. It is, seemingly, a very significant problem to be considered in the case of e.g. data fusion in intelligent traffic control. The intention of the author is to prepare the tools for classification of data which come from various sources. They can be sets (data files) prepared by the user of application, but they can also be one (or many) sets from the UCI machine learning repository (http://archive.ics.uci.edu/ml/).

20

Algebraic separation and shadowing of arbitrary sets

Grzybowski J., Pallaschke D., Urbański R.

Commentationes Mathematicae

|

2013

|

Vol. 53, No. 2

55--64

EN

In this paper we consider a generalization of the separation technique proposed in [10,4,7] for the separation of finitely many compact convex sets Ai, i∈I by another compact convex set S in a locally convex vector space to arbitrary sets in real vector spaces. Then we investigate the notation of shadowing set which is a generalization of the notion of separating set and construct separating sets by means of a generalized Demyanov-difference in locally convex vector spaces.