Informacja

Drogi użytkowniku, aplikacja do prawidłowego działania wymaga obsługi JavaScript. Proszę włącz obsługę JavaScript w Twojej przeglądarce.

English (EN)·Polski (PL)·Yкраїнський (UK)
Przejdź do strony domowej biblioteki Centralna Biblioteka Wojskowa Link otwiera się w nowym oknie Logo biblioteki Prolib Integro - strona głównaCentralna Biblioteka Wojskowa

Wyszukujesz frazę "Classification method" wg kryterium: Temat

  Lista filtrów
Wyrównaj
Wyświetlanie 1-15 z 46
Tytuł:
The attempt to assess the fire risk of non-forest terrestrial ecosystems of Biebrza National Park - A case study
Autorzy:
Szczygieł, Ryszard
Kwiatkowski, Mirosław
Kołakowski, Bartłomiej
Powiązania:
https://bibliotekanauki.pl/articles/2041238.pdf
Data publikacji:
2021
Wydawca:
Instytut Badawczy Leśnictwa
Tematy:
fire risk
non-forest ecosystems
classification method
Opis:
On 19th April, the largest wildfire of the entire history of Biebrza National Park broke out and consumed 5526 ha, mostly overdried grassland with sedges and reeds of Biebrza marshes. The very rapid spread of the fire in the open space, with the blowing wind changing directions and the inaccessibility of the area for fire-fighting vehicles were the main reasons of this third largest wildfire in Poland, after the tragic forest fires in Kuźnia Raciborska (9060 ha burnt) and Potrzebowice (5600 ha burnt) in the memorable year of 1992. After this event, activities were undertaken to develop a fire protection plan for the Biebrza National Park. It took into account an innovative approach during the analysis of the existing fire risk, primarily regarding the non-forest terrestrial ecosystems composed of herbaceous vegetation, which constitutes as much as 61.2% of the park’s area, and to propose protection methods adequate to the threat. The work was completed in the framework of the project entitled ‘Development of the method for assessing the fire risk of non-forest ecosystems and the principles of fire protection for the Biebrza National Park – stage I,’ commissioned for the Forest Research Institute by the Biebrza National Park, financed by the State Forests from the forest fund, in accordance with the contract EZ.0290.1.24.2020. The article presents a preliminary method of classifying the fire risk of non-forest ecosystems, considering the occurrence of the fires in the Biebrza National Park in the years 2007–2020 and the type of vegetation burnt. This method, after supplementary field tests planned in 2021, will enable evaluation of the fire risk, which shall be a premise for the development of a fire protection plan.
Źródło:
Folia Forestalia Polonica. Series A . Forestry; 2021, 63, 2; 167-175
0071-6677
Pojawia się w:
Folia Forestalia Polonica. Series A . Forestry
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Vessel classification method based on vessel behavior in the port of Rotterdam
Autorzy:
Zhou, Y.
Daamen, W.
Vellinga, T.
Hoogendoorn, S.
Powiązania:
https://bibliotekanauki.pl/articles/135299.pdf
Data publikacji:
2015
Wydawca:
Akademia Morska w Szczecinie. Wydawnictwo AMSz
Tematy:
AIS
data analysis
vessel classification
vessel behavior
port
classification method
Opis:
AIS (Automatic Identification System) data have proven to be a valuable source to investigate vessel behavior. The analysis of AIS data provides a possibility to recognize vessel behavior patterns in a waterway area. Furthermore, AIS data can be used to classify vessel behavior into several categories. The analysis results would help the port authority and other equivalent parties in port design and optimization or marine traffic management. For researchers, it provides a systematic way to understand, simulate and predict vessel behavior. This paper focuses on vessel classification in the Botlek area, Rotterdam from the perspective of vessel behavior. In this paper, the vessel properties, including vessel type, GT (Gross Tonnage), length and beam, have been analyzed to investigate the vessel behavior, which is described by four factors including heading, COG (Course over Ground), SOG (Speed over Ground), and position. In order to discover the behavior patterns in normal situations, several thresholds are set in order to filter the collected AIS data to define such situations. By plotting the AIS data, behavioral changes with the changes of properties have been observed. Hence, the correlations between vessel behavior and different vessel properties are investigated. The results reveal that a vessel’s sailing position and COG are both strongly determined by beam, while SOG is affected by GT. For the heading of a vessel, no obvious correlation with any vessel property is found. Each behavioral factor is clustered according to the correlated vessel property. This way, the criteria to classify the vessels are determined. The vessel classification results based on their behavior would likely to lead to more consistency in the analysis, simulation and prediction of the vessel behavior. The reason is that the development of such a simulation model is based on a systematic recognition of the vessel behavior patterns.
Źródło:
Zeszyty Naukowe Akademii Morskiej w Szczecinie; 2015, 42 (114); 86-92
1733-8670
2392-0378
Pojawia się w:
Zeszyty Naukowe Akademii Morskiej w Szczecinie
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Measurement and classification methods using the ASAE S572.1 reference nozzles
Autorzy:
Fritz, B.K.
Hoffmann, W.C.
Czaczyk, Z.
Bagley, W,
Kruger, G.
Henry, R.
Powiązania:
https://bibliotekanauki.pl/articles/66858.pdf
Data publikacji:
2012
Wydawca:
Polska Akademia Nauk. Czytelnia Czasopism PAN
Tematy:
measurement method
classification method
droplet size
laser diffraction
reference nozzle
Opis:
An increasing number of spray nozzle and agrochemical manufacturers are incorporating droplet size measurements into both research and development. Each laboratory invariably has their own sampling setup and procedures. This is particularly true about measurement distance from the nozzle and concurrent airflow velocities. Both have been shown to significantly impact results from laser diffraction instruments. These differences can be overcome through the use of standardized reference nozzles and relative spray classification categories. Sets of references nozzles, which defined a set of classification category thresholds, were evaluated for droplet size under three concurrent air flow velocities (0.7, 3.1 and 6.7 m/s). There were significant, though numerically small, differences in the droplet size data between identical reference nozzles. The resulting droplet size data were used to categorize a number of additional spray nozzles at multiple pressure and air flow velocities. This was done to determine if similar classifications were given across the different airspeeds. Generally, droplet size classifications agreed for all airspeeds, with the few that did not, only differing by one category. When reporting droplet size data, it is critical that data generated from a set of reference nozzles also be presented as a means of providing a relative frame of reference.
Źródło:
Journal of Plant Protection Research; 2012, 52, 4
1427-4345
Pojawia się w:
Journal of Plant Protection Research
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
An overview of classification methods from dermoscopy images in skin lesion diagnostic
Przegląd metod klasyfikacji obrazów dermatoskopowych wykorzystywanych w diagnostyce zmian skórnych
Autorzy:
Michalska, Magdalena
Boyko, Oksana
Powiązania:
https://bibliotekanauki.pl/articles/407654.pdf
Data publikacji:
2020
Wydawca:
Politechnika Lubelska. Wydawnictwo Politechniki Lubelskiej
Tematy:
dermatoscopic images
classification method
neural network
SVM
skin cancer
skin lesion
obraz dermatoskopowy
metoda klasyfikacji
sztuczna sieć neuronowa
nowotwór skóry
zmiany skórne
Opis:
The article contains a review of selected classification methods of dermatoscopic images with human skin lesions, taking into account various stages of dermatological disease. The described algorithms are widely used in the diagnosis of skin lesions, such as artificial neural networks (CNN, DCNN), random forests, SVM, kNN classifier, AdaBoost MC and their modifications. The effectiveness, specificity and accuracy of classifications based on the same data sets were also compared and analyzed.
Artykuł zawiera przegląd wybranych metod klasyfikacji obrazów dermatoskopowych zmian skórnych człowieka z uwzględnieniem różnych etapów choroby dermatologicznej. Opisane algorytmy są szeroko wykorzystywane w diagnostyce zmian skórnych, takie jak sztuczne sieci neuronowe (CNN, DCNN), random forests, SVM, klasyfikator kNN, AdaBoost MC i ich modyfikacje. Porównana i przeanalizowana została również skuteczność, specyficznośc i dokładność klasyfikatów w oparciu o te same zestawy danych.
Źródło:
Informatyka, Automatyka, Pomiary w Gospodarce i Ochronie Środowiska; 2020, 10, 2; 36-39
2083-0157
2391-6761
Pojawia się w:
Informatyka, Automatyka, Pomiary w Gospodarce i Ochronie Środowiska
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Multidimensional Data Classification - Comparison of Isodata and Approximation by Points Methods
Klasyfikacja wielowymiarowych danych - porównanie metody 1 SOD АТА i metody aproksymacji punktami
Autorzy:
Maciuk, Arkadiusz
Powiązania:
https://bibliotekanauki.pl/articles/906292.pdf
Data publikacji:
2009
Wydawca:
Uniwersytet Łódzki. Wydawnictwo Uniwersytetu Łódzkiego
Tematy:
classification
division method
multivalent data
Opis:
Efekt podziału zależy nie tylko od ustalenia kryteriów podziału, ale także od wyboru metody dzielenia. Standardowy algorytm klasyfikacji wielowymiarowych danych ISODATA dzieli wyjściowy zbiór na ustaloną liczbę rozłącznych podzbiorów tak, aby podział ten jak najkorzystniej spełniał przyjęte kryteria. Alternatywą wobec niego jest algorytm oparty na metodzie aproksymacji ustaloną liczbą punktów, którego efektem jest wskazanie obszarów zbioru o dużym stopniu zagęszczenia elementów. Artykuł zawiera porównanie efektów użycia tych metod ze wskazaniem zalet i wad. Omawia też pewne własności klasyfikacji wynikające z konsekwencji wyboru jednej z dwóch omawianych metod.
The effect of division is dependent not only on the criteria of division but also on the chosen method. The standard algorithm of multidimensional data classification ISODATA divides the given set into an assumed number of separable subsets in such a way that the division fulfills best the accepted criteria. An alternative method is approximation by the chosen number of points which in result indicates the areas of a set with large congestion o f the elements. The paper compares effects of using both methods listing their advantages and drawbacks. Apart from presenting the results of division of various sets some characteristics of classification are discussed which are an effect of the choice of one of the above mentioned methods.
Źródło:
Acta Universitatis Lodziensis. Folia Oeconomica; 2009, 228
0208-6018
2353-7663
Pojawia się w:
Acta Universitatis Lodziensis. Folia Oeconomica
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
The effectiveness of corporate bankruptcy models
Skuteczność modeli upadłości przedsiębiorstw
Autorzy:
Kisielińska, J.
Powiązania:
https://bibliotekanauki.pl/articles/2052006.pdf
Data publikacji:
2016
Wydawca:
Akademia Bialska Nauk Stosowanych im. Jana Pawła II w Białej Podlaskiej
Tematy:
renewable energy
classification methods
aggregate method
Opis:
Subject and purpose of work: Research targeted at verification of effectiveness of the selected Polish models of forecasting bankruptcy of an enterprise has been presented within the article. Materials and methods: Within the research financial data from the years 2009-2012 have been used, obtained from 110 companies out of which 55 are companies which underwent bankruptcy and 55 which did not. Discriminant analysis as well as logit models and method of aggregated assessment have been applied. Results: Due to the frequent low quality of classification obtained via single models a method which may be called as aggregated method was suggested, which is based on the assessment of situation of a single unit on indications of the majority of models. This method allowed to increase the correctness of identification up to 87,3%. The share of incorrect identifications is rather high and amounts to 12,7%. Conclusions: The author is of the opinion that the assessment of situation of an entity may not be based solely on the indications of models of bankruptcy prognosis. This assessment may be treated solely indicatively, especially when consequences of a decision made on this basis (i.e. decision on granting a loan) may significantly impact further position of such entity and its surroundings.
Streszczenie Przedmiot i cel pracy: W artykule przedstawiono badania mające na celu weryfikację skuteczności wybranych polskich modeli prognozowania upadłości przedsiębiorstw. Materiały i metody: W badaniach wykorzystano dane finansowe z lat 2009-2012 pochodzące z 110 firm, z czego 55 to firmy, które zbankrutowały i 55, które nie zbankrutowały. Wykorzystano analizę dyskryminacyjną, modele logitowe oraz metodę oceny zagregowanej. Wyniki: Z uwagi na często niską jakość klasyfikacji uzyskaną pojedynczymi modelami zaproponowano metodę, którą nazwać można zagregowaną, opierającą ocenę sytuacji jednostki na wskazaniach większości modeli. Metoda ta pozwoliła podnieść poprawność identyfikacji do 87,3%. Nadal jednak udział błędnych rozpoznań jest dość wysoki i wynosi 12,7%. Wnioski: Autorka uważa, że ocena sytuacji jednostki nie może bazować jedynie na wskazaniach modeli prognozowania upadłości. Ocenę tą można traktować jedynie orientacyjnie zwłaszcza, gdy konsekwencje podjętej na jej podstawie decyzji (np. decyzja o udzieleniu kredytu lub nie) mogą w istotnym stopniu wpływać na dalszą pozycję jednostki lub jej otoczenia.
Źródło:
Economic and Regional Studies; 2016, 09, 1; 5-17
2083-3725
2451-182X
Pojawia się w:
Economic and Regional Studies
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Improving the credibility of the extracted position from a vast collection of job offers with machine learning ensemble methods
Autorzy:
Drozda, Paweł
Ropiak, Krzysztof
Nowak, Bartosz A.
Talun, Arkadiusz
Osowski, Maciej
Powiązania:
https://bibliotekanauki.pl/articles/22615539.pdf
Data publikacji:
2023
Wydawca:
Uniwersytet Warmińsko-Mazurski w Olsztynie
Tematy:
machine learning
web scraping
granularity method
classification
Opis:
The main aim of this paper is to evaluate crawlers collecting the job offers from websites. In particular the research is focused on checking the effectiveness of ensemble machine learning methods for the validity of extracted position from the job ads. Moreover, in order to significantly reduce the training time of the algorithms (Random Forests and XGBoost), granularity methods were also tested to significantly reduce the input training dataset. Both methods achieved satisfactory results in accuracy and F1 measures, which exceeded 96%. In addition, granulation reduced the input dataset by more than 99%, and the results obtained were only slightly worse (accuracy between 1% and 5%, F1 between 3% and 8%). Thus, it can be concluded that the considered methods can be used in the evaluation of job web crawlers.
Źródło:
Technical Sciences / University of Warmia and Mazury in Olsztyn; 2023, 26(1); 125--140
1505-4675
2083-4527
Pojawia się w:
Technical Sciences / University of Warmia and Mazury in Olsztyn
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
A Proposal of New Classification Algorithm
Propozycja nowego algorytmu klasyfikacyjnego
Autorzy:
Korzeniewski, Jerzy
Powiązania:
https://bibliotekanauki.pl/articles/905037.pdf
Data publikacji:
2009
Wydawca:
Uniwersytet Łódzki. Wydawnictwo Uniwersytetu Łódzkiego
Tematy:
classification algorithm
mean shift method
silhouette indices
Opis:
In the paper a new method of classifying points to a predetermined number of classes is presented. The method is based on the use of the sample/window mean shift technique to obtain a synthetic insight into the data set structure. The method's performance is tested on Euclidean space data sets generated by the Milligan's CLUSTGEN programme through comparison with the grouping obtained by the k-means method. The criterion applied are the Rousseeuw's silhouette indices are used as a criterion for comparison.
W artykule przedstawiona jest nowa metoda klasyfikowania punktów zbioru danych do klas, których liczba jest zadana. Metoda oparta jest na wykorzystaniu techniki średniego przesunięcia okna/próby do uzyskania syntetycznego wglądu w strukturę zbioru danych. Działanie metody jest sprawdzone na zbiorach danych z przestrzeni euklidesowych wygenerowanych przy pomocy programu CLUSTGEN poprzez porównanie wyników z grupowaniem uzyskanym metodą k-średnich. Kryterium porównawczym są indeksy sylwetkowe Rousseeuwa.
Źródło:
Acta Universitatis Lodziensis. Folia Oeconomica; 2009, 225
0208-6018
2353-7663
Pojawia się w:
Acta Universitatis Lodziensis. Folia Oeconomica
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Effectiveness of active forgetting in machine learning applied to financial problems
Autorzy:
Nakayama, H.
Yoshii, K.
Powiązania:
https://bibliotekanauki.pl/articles/309245.pdf
Data publikacji:
2002
Wydawca:
Instytut Łączności - Państwowy Instytut Badawczy
Tematy:
pattern classification
potential method
additional learning
forgetting
Opis:
One of main features in financial investment problems is that the situation changes very often over time. Under this circumstance, in particular, it has been observed that additional learning plays an effective role. However, since the rule for classification becomes more and more complex with only additional learning, some appropriate forgetting is also necessary. It seemms natural that many data are forgotten as the time elapses. On the other hand, it is expected more effective to forget unnecessary data actively. In this paper, several methods for active forgetting are suggested. The effectiveness of active forgetting is shown by examples in stock portfolio problems.
Źródło:
Journal of Telecommunications and Information Technology; 2002, 3; 24-29
1509-4553
1899-8852
Pojawia się w:
Journal of Telecommunications and Information Technology
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Acoustic seabed classification applied to Baltic benthic habitat studies: a new approach
Autorzy:
Orlowski, A.
Powiązania:
https://bibliotekanauki.pl/articles/48659.pdf
Data publikacji:
2007
Wydawca:
Polska Akademia Nauk. Instytut Oceanologii PAN
Tematy:
ecology
bottom habitat
acoustic method
acoustic seabed classification
Baltic Sea
bathymetry
acoustic classification
Źródło:
Oceanologia; 2007, 49, 2
0078-3234
Pojawia się w:
Oceanologia
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Effective multi-label classification method with applications to text document categorization
Autorzy:
Glinka, K.
Zakrzewska, D.
Powiązania:
https://bibliotekanauki.pl/articles/94735.pdf
Data publikacji:
2016
Wydawca:
Szkoła Główna Gospodarstwa Wiejskiego w Warszawie. Wydawnictwo Szkoły Głównej Gospodarstwa Wiejskiego w Warszawie
Tematy:
multilabel classification
text categorization
problem transformation method
text management
Opis:
Increasing number of repositories of online documents resulted in growing demand for automatic categorization algorithms. However, in many cases the texts should be assigned to more than one class. In the paper, new multi-label classification algorithm for short documents is considered. The presented problem transformation Labels Chain (LC) algorithm is based on relationship between labels, and consecutively uses result labels as new attributes in the following classification process. The method is validated by experiments conducted on several real text datasets of restaurant reviews, with different number of instances, taking into account such classifiers as kNN, Naive Bayes, SVM and C4.5. The obtained results showed the good performance of the LC method, comparing to the problem transformation methods like Binary Relevance and Label Powerset.
Źródło:
Information Systems in Management; 2016, 5, 1; 24-35
2084-5537
2544-1728
Pojawia się w:
Information Systems in Management
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Spatial differentiation in the social and economic development level in Poland
Autorzy:
Miłek, Dorota
Powiązania:
https://bibliotekanauki.pl/articles/22446539.pdf
Data publikacji:
2018
Wydawca:
Instytut Badań Gospodarczych
Tematy:
region
regional development
regional differentiation
Hellwig’s method
Ward’s method
natural breaks classification
Opis:
Research background: Social and economic development involves a broad spectrum of social, economic and spatial phenomena. The multi-faceted nature of regional development arises directly from the fact that it is shaped by multiple factors. Current discourse emphasises the role of endogenous factors, which indicate the specific nature and the distinctive features of the given territory. Mobilising the endogenous potential ensures stable regional development dynamics. At the moment, one of the fundamental economic problems are the increasing differences in the development of specific regions. Purpose of the article: The purpose of this study is to assess the differentiation of the social and economic level development of Polish Voivodeships, applying the selected assessment methods for the years 2010 and 2015, draw up a rank list of regional units according to their development levels, and identify the groups of Voivodeships sharing similar development levels. The indicators used in this study, characterising the level of the social and economic development, have been systematised according to the following areas: demographics and labour market, regional entrepreneurship, local economy structure, innovation and research & development activities, technical infrastructure, social infrastructure, and the condition and protection of the natural environment. Methods: The level of the social and economic development of Polish Voivodeships was assessed using Zdzisław Hellwig's development pattern method, which made it possible to rank them according to the level of development of Polish Voivodeship. The methodology is supplemented by Ward's agglomerative clustering method, which made it possible to distinguish other Voivodeships according to the analysed phenomenon. The Voivodeship clustering method used Jenks' natural breaks classification method. Findings & Value added: Pursuing the research aims, the authors focused in particular on clear spatial differences. Through the analysis we were able to identify the changes in the social and economic development processes of the Polish regions. The Voivodeships were divided into groups according to their development level: the highest, high, low and the lowest.
Źródło:
Equilibrium. Quarterly Journal of Economics and Economic Policy; 2018, 13, 3; 487-507
1689-765X
2353-3293
Pojawia się w:
Equilibrium. Quarterly Journal of Economics and Economic Policy
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Method vs. form - an attempt to classify cartographic presentation methods
Autorzy:
Mazur, M.
Powiązania:
https://bibliotekanauki.pl/articles/92398.pdf
Data publikacji:
2017
Wydawca:
Oddział Kartograficzny Polskiego Towarzystwa Geograficznego
Tematy:
cartographic methodology
presentation method
presentation form
classification
cartographic presentation methods
Opis:
The author of the article discusses the subject of terminology in the field of cartographic methodology. The general purpose is to propose an internally consistent system of concepts which allows classification of cartographic presentation methods based on unambiguous criteria. For this purpose, in the article the concept of presentation method and the concept of presentation form were differentiated, the cartographic presentation method was divided into stages, and possible data transformation and visualization aspects during application of a method were specified. Then, review of the previous classifications of cartographic presentation methods was conducted which allows differentiation of two fundamental approaches to classification and comparison of the applied criteria. On this basis, the author’s classification of cartographic presentation methods was suggested in which three qualitative and four quantitative methods were differentiated. It constitutes a compromise between accepting unambiguous criteria and the possibility to differentiate methods fixed in cartographic convention.
Źródło:
Polish Cartographical Review; 2017, 49, 2; 47-57
2450-6974
Pojawia się w:
Polish Cartographical Review
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Formalized hierarchically nested expert system for classification of mesic and wet grasslands in Poland
Autorzy:
Kacki, Z.
Swacha, G.
Lengyel, A.
Korzeniak, J.
Powiązania:
https://bibliotekanauki.pl/articles/2130586.pdf
Data publikacji:
2020
Wydawca:
Polskie Towarzystwo Botaniczne
Tematy:
Cocktail method
vegetation classification
Molinio-Arrhenatheretea
syntaxonomy
diversity
Braun-Blanquet approach
Opis:
The goal of this study was to propose a hierarchically nested classification system comprising four principal levels of the Braun-Blanquet system for Polish grasslands of the class Molinio-Arrhenatheretea. Using the Cocktail method, we defined consistent criteria for delimitation of the class, three orders, nine alliances, and 45 associations. Formal definitions were prepared using the summed cover and presence/absence information of species groups and individual dominant species. We created an expert system with a set of assignment rules that unambiguously classify relevés to a single unit at the given abstraction level of the Braun-Blanquet system in such a way that a relevé matched by the definition of a focal vegetation unit must be matched by definitions of all superior units. Of 11,535 relevés classified to Molinio-Arrhenatheretea, 36% were recognized at the association level, and 57% and 85% at the alliance and order level, respectively. All relevés were assigned unambiguously, meaning that a single relevé could not be assigned to more than one unit within the same hierarchical level (no overlap between vegetation units). This study is the first proposal of a hierarchically nested classification system that classifies grassland vegetation at different syntaxonomical levels unequivocally. It is important to create definitions for different syntaxonomical levels because the majority of vegetation patches do not fit to the associations, but can only be assigned to high-rank units such as alliance, order, or class.
Źródło:
Acta Societatis Botanicorum Poloniae; 2020, 89, 4
0001-6977
2083-9480
Pojawia się w:
Acta Societatis Botanicorum Poloniae
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Effect of choice of dissimilarity measure on classification efficiency with nearest neighbor method
Autorzy:
Górecki, Tomasz
Powiązania:
https://bibliotekanauki.pl/articles/729668.pdf
Data publikacji:
2005
Wydawca:
Uniwersytet Zielonogórski. Wydział Matematyki, Informatyki i Ekonometrii
Tematy:
nearest neighbor method
discriminant coordinates
dissimilarity measures
estimators of classification error
Opis:
In this paper we will precisely analyze the nearest neighbor method for different dissimilarity measures, classical and weighed, for which methods of distinguishing were worked out. We will propose looking for weights in the space of discriminant coordinates. Experimental results based on a number of real data sets are presented and analyzed to illustrate the benefits of the proposed methods. As classical dissimilarity measures we will use the Euclidean metric, Manhattan and post office metric. We gave the first two metrics weights and now these measures are not metrics because the triangle inequality does not hold. Howeover, it does not make them useless for the nearest neighbor classification method. Additionally, we will analyze different methods of tie-breaking.
Źródło:
Discussiones Mathematicae Probability and Statistics; 2005, 25, 2; 217-239
1509-9423
Pojawia się w:
Discussiones Mathematicae Probability and Statistics
Dostawca treści:
Biblioteka Nauki
Artykuł
Wyświetlanie 1-15 z 46

Ta witryna wykorzystuje pliki cookies do przechowywania informacji na Twoim komputerze. Pliki cookies stosujemy w celu świadczenia usług na najwyższym poziomie, w tym w sposób dostosowany do indywidualnych potrzeb. Korzystanie z witryny bez zmiany ustawień dotyczących cookies oznacza, że będą one zamieszczane w Twoim komputerze. W każdym momencie możesz dokonać zmiany ustawień dotyczących cookies