|
|
Big data - extraction of key information combining methods of mathematical statistics and machine learning
Masák, Tomáš ; Antoch, Jaromír (advisor)
This thesis is concerned with data analysis, especially with principal component analysis and its sparse modi cation (SPCA), which is NP-hard-to- solve. SPCA problem can be recast into the regression framework in which spar- sity is usually induced with ℓ1-penalty. In the thesis, we propose to use iteratively reweighted ℓ2-penalty instead of the aforementioned ℓ1-approach. We compare the resulting algorithm with several well-known approaches to SPCA using both simulation study and interesting practical example in which we analyze voting re- cords of the Parliament of the Czech Republic. We show experimentally that the proposed algorithm outperforms the other considered algorithms. We also prove convergence of both the proposed algorithm and the original regression-based approach to PCA. vi
|
|
|
Big data - extraction of key information combining methods of mathematical statistics and machine learning
Masák, Tomáš ; Antoch, Jaromír (advisor)
This thesis is concerned with data analysis, especially with principal component analysis and its sparse modi cation (SPCA), which is NP-hard-to- solve. SPCA problem can be recast into the regression framework in which spar- sity is usually induced with ℓ1-penalty. In the thesis, we propose to use iteratively reweighted ℓ2-penalty instead of the aforementioned ℓ1-approach. We compare the resulting algorithm with several well-known approaches to SPCA using both simulation study and interesting practical example in which we analyze voting re- cords of the Parliament of the Czech Republic. We show experimentally that the proposed algorithm outperforms the other considered algorithms. We also prove convergence of both the proposed algorithm and the original regression-based approach to PCA. vi
|
|
|
Big data - extraction of key information combining methods of mathematical statistics and machine learning
Masák, Tomáš ; Antoch, Jaromír (advisor) ; Maciak, Matúš (referee)
This thesis is concerned with data analysis, especially with principal component analysis and its sparse modi cation (SPCA), which is NP-hard-to- solve. SPCA problem can be recast into the regression framework in which spar- sity is usually induced with ℓ1-penalty. In the thesis, we propose to use iteratively reweighted ℓ2-penalty instead of the aforementioned ℓ1-approach. We compare the resulting algorithm with several well-known approaches to SPCA using both simulation study and interesting practical example in which we analyze voting re- cords of the Parliament of the Czech Republic. We show experimentally that the proposed algorithm outperforms the other considered algorithms. We also prove convergence of both the proposed algorithm and the original regression-based approach to PCA. vi
|
|
|
Fault tree analysis
Masák, Tomáš ; Antoch, Jaromír (advisor) ; Bejda, Přemysl (referee)
In the thesis, selected procedures of fault tree analysis and their applications to system reliability analysis are described. Emphasis is placed especially on effective determination of minimal cutsets, provided coherency or monotonicity of studied fault trees. Classical ways of identifying minimal cuts and modern approaches using binary decision diagrams are presented. Precise construction of the theory of binary decision diagrams from the basic theory of Boolean algebra is covered in detail. Algorithms for finding minimal cutsets using binary decision diagrams are at first described theoretically and then they are implemented in C++ programming language. Powered by TCPDF (www.tcpdf.org)
|
|
|
Phones Czech consumers expensively? Market analysis of mobile operators in the Czech Republic.
Masák, Tomáš ; Hudík, Marek (advisor) ; Zemplinerová, Alena (referee)
This thesis will examine whether the price level on the market of mobile services in comparison with other countries is high or not. After performing own international price comparison which revealed high price level in both analyzed consumer baskets thesis offers some possible options which could lead to the current high price level on the market. The observed high market concentration is one of the possible solutions and leads to discover several solutions whose analysis shows that high price level in the Czech mobile services market cannot be associated with only one factor. The price level is mainly affected by the absence of more operators which can be partly attributed to the insufficient steps of market regulator - ČTÚ especially in the area of termination fees and lease licenses to operating in mobile networks. Both of these steps are also very important for the future development of the market.
|
|
|
Leaves F1 from Europe because of tobacco regulation or it will leave as well?
Masák, Tomáš ; Bartoň, Petr (advisor) ; Zajíček, Miroslav (referee)
Bernie Ecclestone (main person of F1 ) said that because of the ban on tobacco advertising F1 leaves from Europe. In today's globalized world is generally survival of World series in Europe unlikely. This thesis investigates how Bernie Ecclestone's statement aboves other economic reasons for leaving Europe. The role of tobacco was critical for F1 in terms of team's sponsorship and F1 itself. In both areas was the tobacco an important element for decades. Describing relations existing in the world of F1 which includes criteria that allow insight into the decisions about the allocation of F1 circuits and there will be a breakdown of the model into two parts. F1's promoter who as head of the body decides on the allocation of circuits direct and F1's teams that has an indirect influence on promoter's decision that derives from their status -- complements of F1's championship. Analysis of decision making model indicates that teams didn't have an effect on a decision of the promoter. The result is that the F1 promoter's decision was not influenced by tobacco advertising but rather revenues from television rights and race sanction fees.
|
guest ::
