|
|
Relational Modeling of Biological Data
Krejčík, Roman ; Toman, Kamil (referee) ; Pokorný, Jaroslav (advisor)
Some applications of data processing in database systems work with hierarchically structured data. Information of such kind is often present in biological data, but it can be also found in another scope. An efficient execution of queries specific for hierarchy needs proper representation of graph structures in a relational database. This thesis deals with a description of known methods for storing of hierarchical data into a relational database. It aims particularly to distinct properties of separate methods and examines fitness of usage for various hierarchies. The thesis includes experiments with real collections of biological data, which are used to compare effectiveness of particular methods in practical way.
|
|
|
Querying Data Streams
Roštár, Matúš ; Ulrych, Jan (referee) ; Pokorný, Jaroslav (advisor)
Recently a new class of data-intensive applications has become widely recognized: application in which the data is modeled best not as persistent relations but rather as transient data streams. In this thesis we address to the review of known data stream management systems, and present our own system. The purpose of this thesis is to expose our new algorithm, based on clustering, sliding window and histogram, that processes data and evaluates continuous queries. We also give experiments over real data to prove that our algorithm is more accurate than other algorithms tested out in the past.
|
|
|
Analytical Processing of XML documents
Melichárek, Petr ; Toman, Kamil (referee) ; Pokorný, Jaroslav (advisor)
In this diploma thesis we present XQuery analytic extension containing clauses well known from SQL language - group by, roll up and cube. In addition we introduce new clause, called topological roll up. The native XML database eXist is an environment for the extension implementation. We add a module, containing earlier mentioned clauses, to the eXist database. We can use implemented clauses as user de ned functions in XQuery language for querying database's documents. Tests have proved that functions are useful in practice. Even for large XML documents and complex queries.
|
|
|
Prolactin And Systemic lupus erythematosus
Moszkorzová, Ludmila ; Dostál, Ctibor (advisor) ; Pokorný, Jaroslav (referee) ; Hrnčíř, Zbyněk (referee)
Prolactin (PRL) is a polypeptide hormone of 23 kDa molecular weight made up of 199 amino acids, and produced by lactotropes, acidophilic cells of the anterior lobe of the pituitary. PRL is also synthetized in some other parts of the brain and in certain peripheral blood elements. Whether this extra-pituitary PRL, also known as PRL-like hormone, interferes with serum PRL radioimmunoassay (RIA) and whether it also has a feedback effect on PRL secretion in the pituitary, has yet to be elucidated. There is, however, proof of its apocrine and paracrine function of cellular growth factor, a function enhancing mitogenesis and lymphocyte differentiation at the site of inflammation and thereby their own production of yet other mediators and immunomodulators - including interleukines (IL) and growth factors. PRL also directly interferes with the synthesis of some acute-phase proteins in the liver (stimulating, e.g., alpha-2-macroglobulin synthesis). As a result of these discoveries, PRL was classed among immunomodulators, and the hypothesis was advanced of its part in the pathogenesis of autoimmune diseases. The aim of the thesis was to verify the presence of hyperprolactinemia (hyper- PRL) in SLE patients compare to patients with other auto-immune diseases and to healthy controls and to find its association...
|
|
|
Nativní indexování rozsáhlých XML databází
Bartoš, Tomáš ; Pokorný, Jaroslav (referee) ; Skopal, Tomáš (advisor)
In the present work we study the indexing methods for large XML databases and their time efficiency when evaluating path queries. There are several ways of indexing XML data but we focus on indexing root- to-leaf paths and grouping them according to the common criteria, path labels. We study the existing methods and combine them in order to create the iXUPT, a novel native indexing concept using path templates, which leverages advantages of current approaches. We provide two variations of our solution depending on the way of handling ancestor-descendant relationships. The first one uses the proposed numbering scheme, while the second one relies on the Rho-Index structure. Furthermore, we prove the feasibility of our concept by the implemented prototype and by evaluating sample regular path expressions represented by XPath queries. We compare the variations between each other and also with other solutions.
|
|
|
Syllable-Based Compression
Lánský, Jan ; Pokorný, Jaroslav (advisor)
Classic textual compression methods work over the alphabet of characters or alphabet of words. For languages with rich morphology as well as for compression of smaller files it can be advantageous to use an alphabet of syllables. For some compression methods like the ones based on Burrows-Wheeler transformation the syllable is a reasonable solution also for large files - even for languages having quite simple morphology. Although the main goal of our research is the compression over the alphabet of syllables, all implemented methods can compress also over the alphabet of words. For small files we use the LZW method and Huffman coding. These methods were improved by the use of initialized dictionary containing characteristic syllables specific for given language. For the compression of very large files we implemented the project XBW allowing combination of compression methods BWT, MTF, RLE, PPM, LZC, and LZSS. We have also tried to compress XML files that are not well-formed. When compressing over a large alphabet, it is necessary to compress also the used alphabet. We have proposed two solutions. The first one works well especially for small documents. We initialize the compression method with a set of characteristic syllables whereas other syllables are coded when necessary character by character. The...
|
|
|
C-store: storing relational data by columns
Částek, Petr ; Žemlička, Michal (referee) ; Pokorný, Jaroslav (advisor)
Non standard approach to the implementation of relational databases this work implements a database system storing the data by columns instead of rows. It is based on the C Store system design and is limited only to the fundamental solution issues. Two areas are being solved: the initial creation of the database, including import of static database data from another source, and querying. The database does not store the data in tables but in projections groups of columns of the same length. By using this database layout it is possible to evaluate queries faster than with database storing by rows as the database engine does not have to read irrelevant columns. Querying is allowed by providing an execution plan consisting of the C Store query operators that work with columns or projections. Composition of these operators can replace queries in SQL notation. This work sets the rules of behavior of various operators, implements these operators and also mentions the optimization of the execution plan. The optimization also includes a C Store system extension providing direct access to data items, which in some cases radically reduce the amount of IO operations. The practical output of the work is a library that provides a programming interface for database access and manipulation.
|
|
|
Representing XML data by C-store
Částková, Zuzana ; Holubová, Irena (referee) ; Pokorný, Jaroslav (advisor)
C-store environment - relational database storing records on the disk by columns. Suitable for aggregations, optimalized for reading. Can it be effectively used as XML database? This thesis considers XML data with and without a schema. Database model is generated from a XML file for both cases. A measure of the model suitability is the possibility of making quick XPath queries. Ancestor, child, younger sibling - these are just some of the discussed XPath axes. Low level system enabling the estimation of number of jumps needed on drive is characteristic for algorithms for each axis. Result? Schema for XML data does not seem useful. Regarding the data without a schema, the use of C-store seems to be optimal. But ... Addition of certain functionality to the system would also be appropriate here - direct access to elements, possibility of not reading the whole column. The thesis extends the definition of C-store with these principles, and adds experiments with real data. The conclusion provides an example implementation of selected algorithms that might be used as guidance for further implementation.
|
|
| |
|
| |
guest ::
RSS feed.