|
Data Mining in Small Business
Sabovčik, František ; Burgetová, Ivana (referee) ; Zendulka, Jaroslav (advisor)
Tato práce si klade za cíl vyhodnotit techniky získávání znalostí pro využití v prostředí malého podnikání. Po prozkoumání dat a konzultace s doménovymi experty byly vybrány dvě úlohy: analyza nákupního košíku a predikce prodejů. Pro analyzu nákupního košíku byl využit algoritmus Relim pro vyhledávání častych itemsetů a metriky určující zajímavost asociačních pravidel. Pro úlohu predikce prodejů byl implementován dekompoziční model, SARIMA, MARS a neuronové sítě s časovym oknem. Modely byly vyhodnoceny. Pomocí optimalizace hyper-parametrů bylo dosaženo přijatelnych vysledků. Oproti předpokladům nedošlo při dodání dat o počasí a využití nelineárních modelů ke zlepšení oproti SARIMA. Predikce byla implementována jako služba na straně serveru pro testování v produkčním prostředí.
|
|
Analysis of Economic Indicators Using Statistical Methods
Wahed, Sam ; Doubravský, Karel (referee) ; Novotná, Veronika (advisor)
The Bachelor’s Thesis analyses through statistical and demographic methods the structure of the population of Czech Republic in a past representation and future projection, in relation to its workforce. The methods of processing the time series and the demographic relations are presented in the theoretical part and these methods are further elaborated in the practical part. The results of the Bachelor’s Thesis that have been discovered by the analysis are assessed in the Conclusion that also contains a proposition of a solution as a contribution to the situation ascertained. The Attachment of the Bachelor’s thesis is a statistical calculator programmed in VBA to facilitate the calculations.
|
|
Multi-Agent System for the Prediction of the Effect of Mutations on Protein Stability
Doseděl, Ondřej ; Martínek, Tomáš (referee) ; Musil, Miloš (advisor)
Proteiny jsou základním stavebním blokem všech žijících organismů, kde jsou zodpovědné za mnoho důležitých procesů. Jsou složeny z řetězců aminokyselin. Tyto řetězce mohou být jakkoliv změněné. Tomuto procesu se říká mutace a může být samovolná nebo indukovaná v laboratoři. Cílem této práce bylo vytvoření nových modelů pro určení stability proteinů. Skládá se ze dvou modelů. První model je multi-agentní systém pro klasifikaci stability proteinů. Nejlepší multi-agentní systém získal přesnost 0.7 a 0.41 MCC. Druhá část se~zabývala predikcí konkrétních hodnot G, kde byl vytvořený Extreme Gradient Boosting model, který získal 1.67 RMSE a 0.53 PCC. Součástí této práce byly představené 2 datasety, které jsou na sobě plně nezávislé, použitelné pro trénování a validaci modelů.
|
|
Interactive software tools for teaching signal processing
Pacas, Ondrej ; Rajmic, Pavel (referee) ; Mangová, Marie (advisor)
This thesis deals with creation of four interactive applications for educational purposes in the field of digital signal processing. The goal of this work is to create four applications which will visually interpretate each of the methods of signal processing. This involves applications for linear regression and least squares method, interpolation and signal reconstruction from its samples, discrete linear convolution and discrete cross-correlation. Applications are created using JavaScript programming language.
|
| |
|
Genetic Programming in Prediction Tasks
Machač, Michal ; Mrázek, Vojtěch (referee) ; Sekanina, Lukáš (advisor)
This thesis introduces various machine learning algorithms which can be used in prediction tasks based on regression. Tree genetic programming and linear genetic programming are explained more thoroughly. Selected machine learning algorithms (linear regression, random forest, multilayer perceptron and tree genetic programming) are compared on publicly available datasets with the use of scikit-learn and gplearn libraries. A core part of this project is a new implementation of linear genetic programming which was developed in C++, tested on common symbolic regression problems and then evaluated on real datasets. Results obtained with the proposed system are compared with the results obtained with gplearn.
|
|
Assessment of the strength of concrete structures by a combination of non-destructive and destructive methods
Masařík, Dominik ; Kocáb, Dalibor (referee) ; Cikrle, Petr (advisor)
This bachelor's thesis is about a combination of destructive and non-destructive methods, including a description of the most modern devices of reflection hardness testers. Further on, an analysis of the new norm ČSN EN 13791 is performed. The practical part of the thesis is about measuring on concrete bodies using the latest rebound hardness tester SilverSchmidt, ultrasonic pulse method and tests of compressive strenght of concrete on individual bodies. The evaluation took place on the base of established linear and quadratic regressions on the test specimens. In situ compressive strength was deternibed according to the procedures specified in the ČSN EN 13791 standard.
|
|
Integration, Visualization, and Mining from Data of World Countries
Dušek, Vladimír ; Rychlý, Marek (referee) ; Bartík, Vladimír (advisor)
This thesis explores the utilization of open data about countries around the world, particularly data in the areas of progress and quality of life. The goal was to design and implement a web application to present this data and further use the data for data mining. The integration and processing of data from open data sources were accomplished using the Apache Airflow platform. The Python framework FastAPI was used to create the API and the JavaScript library ReactJS was used to implement the web application. In the application, the indicators are categorized. Each of them can be displayed for different groups of countries, for different time periods, and in several visualizations. From the domain of data mining, clustering of countries based on a group of indicators and prediction of future development of selected indicators using regression analysis was performed. The final application is available at jakjsmenatom.cz.
|
|
A Library for Convolutional Neural Network Design
Rek, Petr ; Mrázek, Vojtěch (referee) ; Sekanina, Lukáš (advisor)
In this diploma thesis, the reader is introduced to artificial neural networks and convolutional neural networks. Based on that, the design and implementation of a new library for convolutional neural networks is described. The library is then evaluated on widely used datasets and compared to other publicly available libraries. The added benefit of the library, that makes it unique, is its independence on data types. Each layer may contain up to three independent data types - for weights, for inference and for training. For the purpose of evaluating this feature, a data type with fixed point representation is also part of the library. The effects of this representation on trained net accuracy are put to a test.
|
|
Correlations of parameters of selected rock types based on laboratory tests
Závacký, Martin ; Lahuta, Hynek (referee) ; Durmeková,, Tatiana (referee) ; Horák, Vladislav (advisor)
The dissertation thesis deals with the properties of hard rocks important for designing geotechnical structures, such as strength and deformation characteristics, as well as rock failure criteria. This work examines the possibility of using correlation relations to estimate the strength characteristics of rocks from index tests, which can make the rock testing process more efficient. The author described selected laboratory tests of rocks and identified several limitations of the tests procedures based on his own practical experience. The correlation analysis of an extensive data set and the derivation of regression relations for selected dependencies were performed. Furthermore, the rock strength estimation quality of the newly derived regressions was compared with the already published regressions. The analysis shows that the achieved degree of correlation is not sufficient to generalize the examined regressions. A significant reason of the low degree of correlation is the combination of the variability of rock properties and limitations of practical testing procedures. Thus, focus should be paid on calibration of the regression relationships within smaller areas in order to precisely estimate the rock properties as reliable input to the geotechnical design.
|