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

Znaleziono wyników: 38

Liczba wyników na stronie
first rewind previous Strona / 2 next fast forward last
Wyniki wyszukiwania
Wyszukiwano:
w słowach kluczowych:  discriminant analysis
help Sortuj według:

help Ogranicz wyniki do:
first rewind previous Strona / 2 next fast forward last
EN
This paper is centred on a binary classification problem in which it is desired to assign a new object with multivariate features to one of two distinct populations as based on historical sets of samples from two populations. A linear discriminant analysis framework has been proposed, called the minimised sum of deviations by proportion (MSDP) to model the binary classification problem. In the MSDP formulation, the sum of the proportion of exterior deviations is minimised subject to the group separation constraints, the normalisation constraint, the upper bound constraints on proportions of exterior deviations and the sign unrestriction vis-à-vis the non-negativity constraints. The two-phase method in linear programming is adopted as a solution technique to generate the discriminant function. The decision rule on group-membership prediction is constructed using the apparent error rate. The performance of the MSDP has been compared with some existing linear discriminant models using a previously published dataset on road casualties. The MSDP model was more promising and well suited for the imbalanced dataset on road casualties.
EN
Polymer composites are the materials that can be successfully used in the places where high mechanical strength and chemical resistance as well as low absorbability are required. These unique features of polymer composites are obtained mainly due to a suitably selected binder, i.e. a synthetic resin. At the same time, this component accounts for the high production costs of these materials. Partial substitution of the resin with glycolisates obtained using poly(ethylene terephthalate) waste (PET), helps reduce the price of polymeric mortars, while maintaining favourable physicomechanical properties. This modification method also has a beneficial effect on the environment, as it allows the utilisation of a very common waste, which is difficult to dispose of. The article concerns three types of resin mortars, i.e. epoxy, polyester and polyester with the addition of colloidal silica, modified with PET glycolisate. On the basis of the obtained data set and database knowledge mining techniques, such as discriminant analysis and decision trees, it was shown to what extent the type of resin and the presence of an added modifier differentiate the mortar properties. The results obtained with both methods were compared. It was confirmed that these techniques are effective both in the classification and prediction of the type (selection) of mortar in the process of designing new composites.
EN
In this paper, the applications of the multivariate data analysis and optimization on vibration signals from compressors have been tested on the assembly line to identify nonconforming products. The multivariate analysis has wide applicability in the optimization of weather forecasting, agricultural experiments, or, as in this case study, in quality control. The techniques of discriminant analysis and linear program were used to solve the problem. The acceleration and velocity signals used in this work were measured in twenty-five rotating compressors, of which eleven were classified as good baseline compressors and fourteen with manufacturing defects by the specialists in the final acoustic test of the production line. The results obtained with the discriminant analysis separated the conforming and nonconforming groups with a significance level of 0.01, which validated the proposed methodology.
EN
At Lawspet area in Puducherry, India, a unique situation of co-disposal of solid waste dumping and secondary wastewater disposal on land, prevails simultaneously within the same campus. So an attempt is made to assess the combined effect of this co-disposal on the environmental quality and pollution effects on groundwater quality with a view to correctly monitor the situation. Multivariate statistical analysis like hierarchical cluster analysis (HCA) and discriminant analysis (DA) were employed. HCA was performed on borewells, physiochemical parameters and seasons. Borewell clustering identified four clusters illustrating varying degree of groundwater contamination. In parameter clustering, two major clusters were formed indicating hardness and anthropogenic components. Temporal clustering identified three major clusters indicating pre-monsoon, monsoon and post-monsoon. Discriminant analysis revealed nine significant parameters which discriminate four clusters qualitatively affording 86% correct assignation to discriminate among the clusters. Also three major components viz. anthropogenic, hardness and geogenic responsible for groundwater quality in the study area were identified. Conclusively the investigation revealed that the direction of the contaminant transport is towards the southeast direction of the study area, where all the borewells (100%) are affected.
EN
Ultrasound imaging is widely used for breast lesion differentiation. In this paper we propose a neural transfer learning method for breast lesion classification in ultrasound. As reported in several papers, the content and the style of a particular image can be separated with a convolutional neural network. The style, coded by the Gram matrix, can be used to perform neural transfer of artistic style. In this paper we extract the neural style representations of malignant and benign breast lesions using the VGG19 neural network. Next, the Fisher discriminant analysis is used to separate those neural style representations and perform classification. The proposed approach achieves good classification performance (AUC of 0.847). Our method is compared with another transfer learning technique based on extracting pooling layer features (AUC of 0.826). Moreover, we apply the Fisher discriminant analysis to differentiate breast lesions using ultrasound images (AUC of 0.758). Additionally, we extract the eigenimages related to malignant and benign breast lesions and show that these eigenimages present features commonly associated with lesion type, such as contour attributes or shadowing. The proposed techniques may be useful for the researchers interested in ultrasound breast lesion characterization.
EN
The Malacca River basin experienced river water pollution which caused a major deterioration to the ecosystems and environmental health. This study is carried out to assess the water quality data and identify the pattern of water pollution sources in the study area, and also to develop a predictive performance of water quality in the Malacca River basin. A chemometric approach using a combination of HCA, DA, PCA, and MLR, was applied into twenty water quality variables from nine sampling stations that were collected from January until December of 2015 in the river basin. HCA pointed out three clusters, namely Cluster 1 (C1) with low pollution source, Cluster 2 (C2) with moderate pollution source, and Cluster 3 (C3) with high pollution source. In the DA analysis, the results showed 21 variables, 12 variables, and 9 variables for standard mode, forward stepwise mode, and backward stepwise mode, respectively. Meanwhile, the PCA indicated that the main source of pollutants is detected from residential, industrial, commercial, agricultural, animal livestock, as well as forest land. Among the three models developed from MLR analysis, C3 with a high pollution source is detected to be the most suitable model to be used for the prediction of Water Quality Index in the Malacca River basin. This study proposed for an effective river water quality management by having new water quality monitoring network to be designed for more practical use in order to reduce time and effort, as well as cost saving purposes.
PL
Osoby podejmujące decyzje w podmiotach opieki zdrowotnej, poszukując najefektywniejszych sposobów zarządzania, zwracają niemałą uwagę na wykorzystywanie metod statystycznych i wysnuwanie na ich podstawie wniosków. Zasadniczym celem artykułu jest prezentacja zastosowania analizy dyskryminacyjnej do oceny zdolności dyskryminacyjnych zmiennych określających efektywność techniczną szpitali w powiatach i miastach na prawach powiatu województwa śląskiego w roku 2015 oraz przedstawienie krótkiego opisu tej analizy. Przeprowadzone badania wskazały, że decydenci, chcąc poprawić wykorzystanie zasobów, przede wszystkim, muszą zwracać uwagę na osobodni leczenia, a więc czynnik głównie generujący koszty. Takie podejście może sprzyjać poprawie zarządzania środkami finansowymi i rzeczowymi szpitala.
EN
Decision makers in health care units, seeking the most effective management practices, pay close attention to using statistical methods and drawing conclusions from them. The main purpose of the article is to present the application of discriminant analysis to assessing the discriminating abilities of variables determining the technical effectiveness of hospitals in districts and cities with district status of Silesia Province in the year 2015 and to present a short description of this analysis. The study carried out recently showed that decision makers, in order to improve the utilization of resources, must, first of all, pay attention to the bed-days, which is the factor mainly generating costs. This approach can help improve the management of the hospital's financial and material resources.
PL
W celu ograniczenia ryzyka powstawania wysokich stężeń trihalometanów (THM) w wodzie przeznaczonej do spożycia przez ludzi bardzo ważne jest bieżące dostosowywanie warunków operacyjnych eksploatacji układów technologicznych uzdatniania wody do zmieniającej się jakości wody w źródle zasilania. Właściwości rakotwórcze i mutagenne THM-ów w istotny sposób wpływają na częstość monitoringu ich stężeń w systemach zaopatrzenia w wodę, kształtującego znacząco koszty sprzedaży wody. W związku z tym zakres kontroli jakości wody w systemie dystrybucji jest ograniczany do niezbędnego minimum. W takich uwarunkowaniach rośnie znaczenie wykorzystania dobrych modeli matematycznych do symulacji stężeń THM-ów w zmiennych warunkach eksploatacji systemów wodociągowych. W artykule przedstawiono wyniki analizy wpływu różnych czynników na wielkość stężenia generowanych THM-ów, takich jak: odczyn wody pH, absorbancja UV w 272 nm, utlenialność, OWO, dawka chloru, stężenie chloru pozostałego. Wszystkie dane zgromadzono podczas badań prowadzonych w rzeczywistym systemie zaopatrzenia w wodę. Celem analiz było zastosowanie narzędzi statystycznych do określenia, które z wymienionych czynników (zmienne niezależne) mają największy wpływ na zmienną zależną, czyli wielkość powstających THM-ów w sieci wodociągowej.
EN
It is very important to adapt water treatment arrangements to changing conditions of raw water quality in the order to reduce the risk of too high concentration of generated trihalomethanes (THMs). Since THMs are danger for human health because of their mutagenic and carcinogenic character, their monitoring in whole water supply system should be frequent. The precise monitoring of THM is expensive so it is limited to indispensable minimal range because of its close relation to price of water delivered to consumers. Hence increases the role of good mathematical models predicting the concentration of THMs in changing operating condition of real water supply system. In this paper authors analyzed different factors influencing the THMs concentration, such as pH, temperature, UV absorbance 272, chemical oxygen demand, total organic carbon, chlorine dose, residual chlorine. All data were collected in real water supply system. The statistical tools were used to identify which of listed factors (independent variables) are of the most impact on dependent variable (level of THMs generated in water pipe network).
9
EN
Dimension reduction and feature selection are fundamental tools for machine learning and data mining. Most existing methods, however, assume that objects are represented by a single vectorial descriptor. In reality, some description methods assign unordered sets or graphs of vectors to a single object, where each vector is assumed to have the same number of dimensions, but is drawn from a different probability distribution. Moreover, some applications (such as pose estimation) may require the recognition of individual vectors (nodes) of an object. In such cases it is essential that the nodes within a single object remain distinguishable after dimension reduction. In this paper we propose new discriminant analysis methods that are able to satisfy two criteria at the same time: separating between classes and between the nodes of an object instance. We analyze and evaluate our methods on several different synthetic and real-world datasets.
10
Content available remote Classification of falling asleep states using HRV analysis
EN
The article presents the results of studies on drowsiness and drowsiness detection performed using heart rate variability analysis (HRV). The results of those studies indicate that the most significant parameters, from the standpoint of classification of drowsiness are the following parameters of the HRV analysis: the low and high frequency band the ratio of the tachogram power in the LF and HF bands, and the total power distribution. The best detection results were obtained for the following methods, in the following order: the nearest neighborhood with metrics: standardized Euclides and Mahalanobis, the square discriminant analysis, and the Bayesian classifier. In order to classify drowsiness periods, a neural network was also used; it consisted of four inputs, six neurons in the hidden layer, and three outputs, one of which was assigned to one of the accepted classes. In order to obtain the most effective learning, a linear feed forward network was designed using back propagation of errors and the RPROP algorithm. In the case of this type of networks, the achieved accuracy of the individual classes was on the level of 98.7%.
EN
Proper characteristics of the traffic flow is a particularly important issue in the process of optimizing the efficiency of transport networks as well as in the traffic control systems. One of the elementary parameters of traffic flows is the structure of vehicles, which evaluation, in case of the automatic systems, requires the implementation of proper algorithms and methods for vehicle classification. In the paper is presented a method of vehicles classification using the discriminant analysis. Furthermore authors developed a classifier, which aggregate data according to classification 8+1 in accordance with the TLS specifications and according with the classification presented in the specification COST 323. As input dataset to the classification method were used vehicle parameters recorded by the weight in motion systems.
EN
In the study, environmetric methods were successfully performed a) to explore natural and anthropogenic controls on reservoir water quality, b) to investigate spatial and temporal differences in quality, and c) to determine quality variables discriminating three reservoirs in Izmir, Turkey. Results showed that overall water quality was mainly governed by “natural factors” in the whole region. A parameter that was the most important in contributing to water quality variation for one reservoir was not important for another. Between summer and winter periods, difference in arsenic concentrations were statistically significant in the Tahtalı, Ürkmez and iron concentrations were in the Balçova reservoirs. Observation of high/low levels in two seasons was explained by different processes as for instance, dilution from runoff at times of high flow seeped through soil and entered the river along with the rainwater run-off and adsorption. Three variables “boron, arsenic and sulphate” discriminated quality among Balçova & Tahtalı, Balçova & Ürkmez and two variables “zinc and arsenic” among the Tahtalı & Ürkmez reservoirs. The results illustrated the usefulness of multivariate statistical techniques to fingerprint pollution sources and investigate temporal/spatial variations in water quality.
13
Content available remote Discriminant analysis of voice commands in a car cabin
EN
Automatic speech recognition systems are used in vehicles. With this application it is possible to control the navigation system, air conditioning system, media player, and make phone calls by using voice commands. The effectiveness of speech recognition systems depends largely on the acoustic conditions in the cabin of the vehicle. Recognition accuracy determines the ability to extend the functionality of such systems beyond the basic functions listed above. The article shows the preliminary results of research on speech recognition and evaluation of speech intelligibility in the vehicle cabin. The purpose of this article is to present the influence of the background noise levels in a car cabin on speech intelligibility, and to investigate the discriminant analysis as a robust classifier for the speech recognition process.
PL
Automatyczne systemy rozpoznawania mowy są stosowane w pojazdach. Dzięki tej aplikacji możliwe jest sterowanie systemem nawigacji, klimatyzacją, odtwarzaczem multimedialnym i wykonywanie połączeń telefonicznych za pomocą poleceń głosowych. Skuteczność systemów rozpoznawania mowy zależy w dużej mierze od warunków akustycznych w kabinie pojazdu. Dokładność rozpoznawania określa zdolność do rozszerzenia funkcjonalności takich systemów poza podstawowe funkcje wymienione powyżej. W pracy przedstawiono wstępne wyniki badań nad rozpoznawaniem mowy i oceną zrozumiałości mowy w kabinie pojazdu. Celem pracy było przedstawienie wpływu poziomu tła w kabinie samochodu na zrozumiałość mowy i zbadanie analizy dyskryminacyjnej jako klasyfikatora w procesie rozpoznawania mowy.
14
Content available remote Analiza archeometryczna ceramiki pradziejowej z zastosowaniem SEM-EDX
PL
Artykuł prezentuje wyniki analizy archeometrycznej przeprowadzonej na kolekcji 50 fragmentów naczyń z wczesnej epoki brązu. Fragmenty pozyskano z sześciu stanowisk archeologicznych położonych na ziemi chełmińskiej: Biały Bór, Grudziądz Mniszek, Małe Radowiska, Toruń Grębocin, Wałyczyk, Zieleń i jednego na ziemi dobrzyńskiej: Skrzypkowo. Wszystkie analizowane materiały należy łączyć ze społecznościami kultury iwieńskiej, których rozwój przypada na przełom III i II tys. p.Ch. Fragmenty naczyń przeprowadzono do postaci proszku ceramicznego i poddano oznaczeniom składu chemicznego metodą dyspersji energii promieniowania rentgenowskiego (EDX) z zastosowaniem skaningowej mikroskopii elektronowej (SEM). Uzyskano spektrogramy matryc chemicznych, na których udało się określić ilościowo udział następujących pierwiastków: C, Na, Mg, Al, Si, P, K, Ca, Ti, Mn, Fe, Cu. Dane pomiarowe zostały poddane analizie archeometrycznej, polegającej m.in. na zastosowaniu metody dyskryminacyjnej. W ten sposób prześledzono reguły asocjowania poszczególnych zbiorów, z uwzględnieniem ich zróżnicowania na stanowiska archeologiczne (z których pozyskano badane fragmenty), grupy technologiczne (do których zaklasyfikowano poszczególne fragmenty) oraz dodatkowo w ramach jednego mikroregionu osadniczego. Wyniki analizy dyskryminacyjnej podały zmienne (zawartości pierwiastków), które najlepiej dyskryminowały zbiór analizowanych fragmentów pod względem ich pochodzenia i przynależności do poszczególnych grup technologicznych. Analiza średniej zawartości Al wskazuje, że utrzymywał się on na porównywalnym poziomie (11-13)% mas. (poza ceramiką ze stanowiska w Białym Borze, gdzie zarejestrowano podwyższony poziom jego zawartości - około 16% mas.). Na tej podstawie stwierdzono, że wytwórcy naczyń na poszczególnych osadach we wczesnej epoce brązu stosowali gliny plastyczne. Natomiast obniżona do (1-2)% mas. średnia zawartość P wskazuje, że nie praktykowano schudzania masy garncarskiej domieszkami pochodzenia roślinnego. Uzyskane wyniki pozwoliły sformułować wnioski na temat stosowania przez społeczności kultury iwieńskiej podobnych źródeł surowcowych, zwłaszcza w obrębie mikroregionu osadniczego Małe Radowiska-Wałyczyk-Zieleń. Ponadto, zaobserwowano tendencję utrzymywania pewnego kanonu technologicznego wytwarzania naczyń, obejmującego podział na ceramikę o charakterze "reprezentacyjnym" i "kuchennym”. Tendencja ta była trwale wpisana w reguły garncarskie i kultywowana niezależnie od miejsca zamieszkiwania i dostępności do złóż surowcowych.
EN
This article presents the results of an archaeometric analysis performed on a collection of 50 Early Bronze Age pottery sherds. The potsherds were obtained from six archaeological sites located on the Chełmno Land: Biały Bór, Grudziądz Mniszek, Małe Radowiska, Toruń Grębocin, Wałyczyk, Zieleń and one on the Dobrzyń Land: Skrzypkowo. All the materials analyzed in the article are connected with the Iwno Culture societies from the turn of the 3rd and 2nd millenium B.C. The potsherds were grinded, and the chemical composition of the resultant powder was identified by Energy-Dispersive X-ray Spectroscopy (EDX) combined with a scanning electron microscopy (SEM). Chemical matrix spectrograms were obtained. The spectrograms allowed performing quantitative determination of the following elements: C, Na, Mg, Al, Si, P, K, Ca, Ti, Mn, Fe, Cu. Data set was statistically elaborated using an archaeometric analysis, including i.a. the discriminant analysis. It permitts to follow an association rules within the individual sets, such as the site (from where the potsherds were excavated), technological groups (which the potsherds were classified into) and additionally within one settlement microregion. The results of the discriminant analysis indicated the variables (elements concentration) that best of all discriminate the set of the analyzed potsherds in terms of their provenance and their belonging to the technological groups. Analysis of mean values of the Al concentration indicates a level of 11-13 wt% (except of the pottery from Biały Bór site where the elevated concentration of Al – amounted to 16 wt% was registered). On this basis, it has been found that the vessels manufacturers from each the Early Bronze Age settlements used plastic clay sources. In contrast, low mean concentration of P (amounted to about 1-2 wt%) indicates that a practice of weakening the clay paste with floral additives did not take place. The obtained results permitted to draw conclusions on the use of similar clay sources by the communities of the Iwno Culture, especially within the settlement microregion Małe Radowiska-Wałyczyk-Zieleń. In addition, the presence of a tendency of cultivating some technological canon within vessels manufacturing, which included differentiation between “representative” and ”kitchen” pottery, was observed. This tendency had been permanently existing in the pottery manufacturing rules and it had been cultivated regardless of a place of habitation, and access to raw material sources.
EN
This study illustrates the benefits of statistical techniques to analyze spatial and temporal variations in water quality. In this scope water quality differentiation caused by anthropogenic and natural factors in the Tahtali and Balçova reservoirs in western Turkey was investigated using discriminant analysis-DA, Mann Whitney U techniques. Effectiveness of pollution prevention measures was analyzed by Mann Kendall and Sen’s Slope estimator methods. The water quality variables were divided into three groups as physical-inorganic, organic and inorganic pollution parameters for the study. Results showed that water quality between reservoirs was differentiated for “physical-inorganic” and “organic pollution” parameters. Degree of influence of water quality by urbanization was higher in the Tahtali reservoir and in general, no trend detection at pollution indicators explained by effective management practices at both sites.
EN
The aim of this study was to compare normal (NL) and tracheoesophageal (TE) vowel speech signals in order to show differences between them. Cepstral features extracted from vowels of NL and TE speech were analyzed using discriminant analysis. The comparison was made on the basis of the classification function coefficients and the results of the classification for each speech. Vowels recordings were acquired from 10 NL speakers and 12 TE speakers. Discriminant analysis was based on cepstral features extracted from vowel recordings, and was performed separately for NL speech and TE speech. Then a comparison between the coefficients of classification functions of NL and TE vowels using the Euclidean distance was made. Based on the resulting classification matrix of NL and TE speech, the results of classification were compared. The discriminant analysis based on cepstral features showed 79% of the mean classification score for TE speech. The Euclidean distance showed low differences between vowel /a/ of NL speech and vowel /a/ of TE speech and between vowel /o/ of NL speech and vowel /o/ of TE speech.
PL
Celem pracy było porównanie sygnału mowy przetokowej (TE) do mowy normalnej (NL) w zakresie samogłosek, aby wykazać różnice pomiędzy sygnałami. Współczynniki cepstralne uzyskane z samogłosek mowy NL i TE poddano analizie dyskryminacyjnej. Na podstawie uzyskanych współczynników funkcji klasyfikacyjnych oraz otrzymanych wyników klasyfikacji dokonano porównania sygnałów mowy NL i TE. Nagrania samogłosek pozyskane zostały od 10 mówców mowy NL i 12 mówców mowy TE. Analizę dyskryminacyjną przeprowadzono w oparciu o współczynniki cepstralne oddzielnie dla mowy NL i mowy TE. Następnie dokonano porównania uzyskanych współczynników funkcji klasyfikacyjnych samogłosek mowy NL i mowy TE, wykorzystując do tego celu odległość Euklidesa. Na podstawie macierzy klasyfikacji otrzymanej dla mowy NL i TE porównano rezultaty klasyfikacji. Analiza dyskryminacyjna w oparciu o współczynniki cepstralne wykazała 79% jako średni wynik klasyfikacji dla mowy TE. Odległość Euklidesa wskazuje na najmniejsze różnice w zakresie samogłoski /a/ i /o/ mowy NL i TE.
17
EN
The aim of this study was applications of cerebrospinal fluid (CSF) NMR-based metabolic fingerprinting to amyotrophic lateral sclerosis (ALS) as possible early diagnostic tool. Two CSF sample categories were collected: 9 ALS patients and 13 age-matched control patients (without neurological disease). Metabolic profile of the CSF was determined by high resolution proton NMR spectroscopy. For statistical analysis magnitudes of 33 signals of the NMR spectrum were selected. Partial least square discriminant analysis (PLS-DA) and orthogonal PLS-DA (OPLS-DA) modeling were used to find potential biomarkers of the disease. Those analyses showed that it was possible to distinguish the ALS patients from the control ones on the basis of the CSF metabolic profile. Significantly higher levels of metabolites observed in the patients with ALS may represent the state of anaerobic metabolism and excitotoxicity.
PL
W artykule przedstawiono sposób oszacowania stanu bezpieczeństwa ruchu drogowego na odcinkach przejść drogowych przez miejscowości za pomocą zmiennej jakościowej typu zabudowy. Zmienna ta pozwala zastąpić wiele zmiennych charakteryzujących miejscowość poprzez tylko jedną zmienną niezależną. W tym celu stworzono typologię form zagospodarowania otoczenia drogi. Statystyczną istotność czynników wpływających na klasyfikowane typy zabudowy wykazano za pomocą analizy dyskryminacyjnej. Aby ocenić wpływ typu zabudowy na bezpieczeństwo ruchu drogowego zbudowano uogólnione modele regresyjne szacowania zdarzeń drogowych. Przedstawione związki typu zabudowy z jej oddziaływaniem na ruch drogowy na drodze nadrzędnej ujęte w postaci zależności funkcyjnych, można wykorzystać w pracach planistycznych w ramach weryfikacji zmian zagospodarowania otoczenia dróg, a także jako narzędzie analiz efektywności projektowanych zmian dostępności do dróg.
EN
The paper presents an assessment of road safety on roads through built-up areas with application of only one qualitative variable (type of development). This variable can replace many of variables describing road surroundings development of a built-up area. Therefore a typology of forms of road surrounding development is proposed. Statistical significance of the factors affecting types of development is classified using discriminant analyses. To assess the impact of development type on road safety, accident prediction models are developed applying generalized regression models. The presented results can be used in the process of land use planning especially in case of road surroundings development, as well as a tool for analyses of the effectiveness of modification in access to roads.
19
Content available Cepstral analysis of vowels of esophageal speakers
EN
The aim of this study was to compare normal (NL) and esophageal (ES) speech signals in scope of vowels in order to show differences between signals. A discriminant analysis based on cepstral features extracted from vowels of NL and ES speech was performed. The comparison was made on the basis of the classification function coefficients and the results of the classification for each speech. Vowels recordings were acquired from 10 NL speakers and 10 ES speakers. The discriminant analysis was based on cepstral features extracted from vowel recordings, and was performed separately for NL speech and ES speech. Then a comparison between coefficients of classification functions of NL and ES vowels using the Euclidean distance was made. Based on the resulting classification matrix of NL and ES speech, the results of classification were compared. The discriminant analysis based on cepstral features showed 76% of the mean classification score for ES speech and 90% for NL speech. The Euclidean distance showed low differences between the vowel /a/ of NL speech and the vowel /a/ of ES speech and between the vowel /e/ of NL speech and the vowel /e/ of ES speech.
PL
Celem pracy było porównanie sygnału mowy normalnej (NL) i przełykowej (ES) w zakresie samogłosek w celu wykazania różnic pomiędzy sygnałami. Przeprowadzono analizę dyskryminacyjną współczynników cepstralnych uzyskanych z samogłosek mowy NL i ES. Porównania dokonano na podstawie uzyskanych współczynników funkcji klasyfikacyjnych oraz otrzymanych wyników klasyfikacji dla każdej mowy. Sygnał mowy każdej samogłoski pozyskany został od 10 mówców mowy NL i 10 mówców mowy ES. Analizę dyskryminacyjną przeprowadzono w oparciu o współczynniki cepstralne oddzielnie dla mowy NL i mowy ES. Następnie dokonano porównania uzyskanych współczynników funkcji klasyfikacyjnych samogłosek mowy NL i mowy ES, wykorzystując do tego celu odległość Euklidesa. Na podstawie macierzy klasyfikacji otrzymanej dla mowy NL i ES porównano rezultaty klasyfikacji. Analiza dyskryminacyjna w oparciu o współczynniki cepstralne wykazała 76% jako średni wynik klasyfikacji dla mowy ES, natomiast 90% dla mowy NL. Odległość Euklidesa wskazuje na najmniejsze różnice w zakresie samogłoski /a/ i /e/ mowy NL i ES.
EN
The purpose of this article is to identify the factors influencing the probability of winning in public procurement procedures and to assess the strength of their impact from the perspective of both: the bidder and procurer. The research was conducted with the use of series of quantitative methods: binary logistic regression, discriminant analysis and cluster analysis. It was based on a sample consisting of public tenders, in which the examined company performed the role of a bidder. Thus, the research process was aimed at both identifying the factors of success and estimating the probability of achieving it, where it was possible to obtain probabilities. The main idea of this research is to answer questions about the utility of various methods of quantitative analysis in the case of analyzing determinants of success. Results of the research are presented in the following sequence of sections: characteristics of the examined material, the process of modelling the probability of winning, evaluation of the quality of the results obtained.
first rewind previous Strona / 2 next fast forward last
JavaScript jest wyłączony w Twojej przeglądarce internetowej. Włącz go, a następnie odśwież stronę, aby móc w pełni z niej korzystać.