No description available.

Over the past decades simulation techniques, and in particular finite element method, have been used successfully to predict the response of systems across a whole range of industries including aerospace, automotive, chemical processes, geotechnical engineering and many others. In these numerical analyses, the behaviour of the actual material is approximated with that of an idealised material that deforms in accordance with some constitutive relationships. Therefore, the choice of an appropriate constitutive model that adequately describes the behaviour of the material plays an important role in the accuracy and reliability of the numerical predictions. During the past decades several constitutive models have been developed for various materials. In recent years, by rapid and effective developments in computational software and hardware, alternative computer aided pattern recognition techniques have been introduced to constitutive modelling of materials. The main idea behind pattern recognition systems such as neural network, fuzzy logic or genetic programming is that they learn adaptively from experience and extract various discriminants, each appropriate for its purpose. In this thesis a novel approach is presented and employed to develop constitutive models for materials in general and soils in particular based on evolutionary polynomial regression (EPR). EPR is a hybrid data mining technique that searches for symbolic structures (representing the behaviour of a system) using genetic algorithm and estimates the constant values by the least squares method. Stress-strain data from experiments are employed to train and develop EPR-based material models. The developed models are compared with some of the existing conventional constitutive material models and its advantages are highlighted. It is also shown that the developed EPR-based material models can be incorporated in finite element (FE) analysis. Different examples are used to verify the developed EPR-based FE model. The results of the EPR-FEM are compared with those of a standard FEM where conventional constitutive models are used to model the material behaviour. These results show that EPR-FEM can be successfully employed to analyse different structural and geotechnical engineering problems.

The thesis deals with the comparison of conventional and heuristic methods in data mining used for binary classification. In the theoretical part, four different models are described. Model classification is demonstrated on simple examples. In the practical part, models are compared on real data. This part also consists of data cleaning, outliers removal, two different transformations and dimension reduction. In the last part methods used to quality testing of models are described.

No description available.

This thesis focuses on comparison of selected data mining methods for solving classification tasks with the method of logistic regression. First part of the thesis briefly introduces data mining as a scientific discipline and classification task is shown in the context of knowledge data discovery. Next part explains the principle of particular methods amongst which, along with logistic regression, artificial neural networks, classification decision trees and Support Vector Machine method were selected. Together with mathematical background of each algorithm, demonstration of how the classification functions for new examples is mentioned. Analytical part of this thesis tests decribed methods on real-world data from the Lending Club company and they are compared based on classification accuracy. Towards the end, an evaluation of logistic regression is made in terms of whether its majority position is due to historical reasons or for its high classification accuracy compared to other methods.

Data mining is a new discipline that occurs with increasing amount of stored data and the increasing need to obtain the information hidden in them. It is focused on the mining of potentially useful information from large data sets and it lies at the intersection of statistics, machine learning, artificial intelligence, databases and other areas. The aim of this thesis is to present the process of data mining with an emphasis on its connection with statistics and to describe a selection of statistical methods widely used in this field and which were also used in the applied data mining problem in this thesis. Real data from purchases in the online store show that using different methods gives different results and interesting information about purchasing behavior, and also proves that not all methods are always applicable to all types of tasks.

No description available.

Predictive power of classification models can be evaluated by various measures. The most popular measures in data mining (DM) are Gini coefficient, Kolmogorov-Smirnov statistic and lift. These measures are each based on a completely different way of calculation. If an analyst is used to one of these measures it can be difficult for him to asses the predictive power of a model evaluated by another measure. The aim of this thesis is to develop a method how to convert one performance measure into another. Even though this thesis focuses mainly on the above-mentioned measures, it deals also with other measures like sensitivity, specificity, total accuracy and area under ROC curve. During development of DM models you may need to work with a sample that is stratified by values of the target variable Y instead of working with the whole population containing millions of observations. If you evaluate a model developed on a stratified data you may need to convert these measures to the whole population. This thesis describes a way, how to carry out this conversion. A software application (CPM) enabling all these conversions makes part of this thesis. With this application you can not only convert one performance measure to another, but you can also convert measures calculated on a stratified sample to the whole population. Besides the above mentioned performance measures (sensitivity, specificity, total accuracy, Gini coefficient, Kolmogorov-Smirnov statistic), CPM will also generate confusion matrix and performance charts (lift chart, gains chart, ROC chart and KS chart). This thesis comprises the user manual to this application as well as the web address where the application can be downloaded. The theory described in this thesis was verified on the real data.