5 resultados para Discriminant

em Doria (National Library of Finland DSpace Services) - National Library of Finland, Finland


Relevância:

10.00% 10.00%

Publicador:

Resumo:

In this thesis author approaches the problem of automated text classification, which is one of basic tasks for building Intelligent Internet Search Agent. The work discusses various approaches to solving sub-problems of automated text classification, such as feature extraction and machine learning on text sources. Author also describes her own multiword approach to feature extraction and pres-ents the results of testing this approach using linear discriminant analysis based classifier, and classifier combining unsupervised learning for etalon extraction with supervised learning using common backpropagation algorithm for multilevel perceptron.

Relevância:

10.00% 10.00%

Publicador:

Resumo:

Työn tavoitteena oli kartoittaa Etelä-Karjalan koulutuskuntayhtymän hankintaprosessin nykytila, analysoida saatuja tuloksia ja pohtia menetelmiä hankintatoimintaan liittyvien prosessien parantamiseksi. Päähuomio kohdistettiin materiaalihankintoihin. Julkisia hankintoja ohjaavat lait ja asetukset, joiden keskeisimmät tavoitteet ovat avoimuus, tasapuolisuus ja syrjimättömyys. Lisäksi laki viranomaisen toiminnan julkisuudesta vaikuttaa asiakirjojen julkisuuteen hankintaprosessin aikana.Tutkimuksessa havaittiin, että erittäin suuri osa ostoista on markkamääräisesti pieniä ostoja, jotka aiheuttavat hintaansa nähden suuret hankintakustannukset. Toimittajia on paljon ja rahallisesti pienten toimittajien määrä on erittäin suuri. Tästä on seurauksena se, että toimittajayhteistyö on useiden toimittajien kanssa hyvin satunnaista ja toimitusten valvonta ja toimittajaseuranta vaikeutuu. Toimittajien kehittämiseen ei tästä johtuen enää riitä resursseja. Toimittajayhteistyö julkisiin hankintoihin erikoistuneen kauppatalon kanssa on vasta alullaan ja yhteistyön määrä vaihtelee paljon eri koulutusyksiköissä. Yllättäviä tuloksia tuotti toimittajan hintojen kehitys sopimuskauden aikana. Hinnoissa oli tapahtunut kaikissa tuoteryhmissä huomattavaa nousua vuoden aikana.Hankintatoimintaa voidaan kehittää nopeuttamalla nykyisiä prosesseja hyväksikäyttäen olemassa olevaa Intranet-tekniikkaa. Samoin sähköinen kaupankäynti on useissa tuoteryhmissä varteenotettava ja tulevaisuudessa edelleen kehittyvä vaihtoehto. Toimittajaseurantaa tulee tehostaa ja toimittajien kanssa tulee pyrkiä avoimeen ja jatkuvaan yhteistyöhön.Hankintayhteistyön kehittäminen omistajakuntien kanssa ja Etelä-Karjalan julkisen sektorin yhteishankintoihin sitoutuminen olisivat tehokas tapa säästää hankintakustannuksissa.

Relevância:

10.00% 10.00%

Publicador:

Resumo:

Dirt counting and dirt particle characterisation of pulp samples is an important part of quality control in pulp and paper production. The need for an automatic image analysis system to consider dirt particle characterisation in various pulp samples is also very critical. However, existent image analysis systems utilise a single threshold to segment the dirt particles in different pulp samples. This limits their precision. Based on evidence, designing an automatic image analysis system that could overcome this deficiency is very useful. In this study, the developed Niblack thresholding method is proposed. The method defines the threshold based on the number of segmented particles. In addition, the Kittler thresholding is utilised. Both of these thresholding methods can determine the dirt count of the different pulp samples accurately as compared to visual inspection and the Digital Optical Measuring and Analysis System (DOMAS). In addition, the minimum resolution needed for acquiring a scanner image is defined. By considering the variation in dirt particle features, the curl shows acceptable difference to discriminate the bark and the fibre bundles in different pulp samples. Three classifiers, called k-Nearest Neighbour, Linear Discriminant Analysis and Multi-layer Perceptron are utilised to categorize the dirt particles. Linear Discriminant Analysis and Multi-layer Perceptron are the most accurate in classifying the segmented dirt particles by the Kittler thresholding with morphological processing. The result shows that the dirt particles are successfully categorized for bark and for fibre bundles.

Relevância:

10.00% 10.00%

Publicador:

Resumo:

Problem of modeling of anaesthesia depth level is studied in this Master Thesis. It applies analysis of EEG signals with nonlinear dynamics theory and further classification of obtained values. The main stages of this study are the following: data preprocessing; calculation of optimal embedding parameters for phase space reconstruction; obtaining reconstructed phase portraits of each EEG signal; formation of the feature set to characterise obtained phase portraits; classification of four different anaesthesia levels basing on previously estimated features. Classification was performed with: Linear and quadratic Discriminant Analysis, k Nearest Neighbours method and online clustering. In addition, this work provides overview of existing approaches to anaesthesia depth monitoring, description of basic concepts of nonlinear dynamics theory used in this Master Thesis and comparative analysis of several different classification methods.

Relevância:

10.00% 10.00%

Publicador:

Resumo:

This master’s thesis studies the probability of bankruptcy of Finnish limited liability companies as a part of credit risk assessment. The main idea of this thesis is to build and test bankruptcy prediction models for Finnish limited liability companies that can be utilized in credit decision making. The data used in this thesis consists of historical financial statements from 2112 Finnish limited liability companies, half of which have filed for bankruptcy. A total of four models are developed, two with logistic regression and two with multivariate discriminant analysis (MDA). The time horizon of the models varies from 1 to 2 years prior to the bankruptcy, and 14 different financial variables are used in the model formation. The results show that the prediction accuracy of the models ranges between 81.7% and 88.9%, and the best prediction accuracy is achieved with the one year prior the bankruptcy logistic regression model. However the difference between the best logistic model and the best MDA model is minimal. Overall based on the results of this thesis it can be concluded that predicting bankruptcy is possible to some extent, but naturally the results are not perfect.