Jiri Dvorsky

On SVD-free latent semantic indexing for image retrieval for application in a hard industrial environment

Retrieval system based on latent semantic indexing (LSI) is presented in our paper. LSI is usually based on single value decomposition, which is time-consuming computation. New, fast algorithm for computing the LSI was developed. The algorithm was tested on text and image data collection. Experimental results are given. The algorithms are intended for practical industrial use within a control and information system operating at coking plant at Nova hut Ostrava. The description of the industrial environment is given.

On maps comparison methods

The mutual similarity of two maps can be most easily compared visually. In this case, the degree of similarity is very subjective. It is therefore necessary to find an objective method for measuring similarity. This chapter presents a method based on Singular Value Decomposition (SVD).

Searching for quasigroups for hash functions with genetic algorithms

In this study we discuss a method for evolution of quasigroups with desired properties based on genetic algorithms. Quasigroups are a well-known combinatorial design equivalent to the more familiar Latin squares. One of their most important properties is that all possible elements of certain quasigroup occur with equal probability. The quasigroups are evolved within a framework of a simple hash function. Prior implementations of quasigroups were based on look-up table of the quasigroup, which is infeasible for large quasigroups. In contrast, analytic quasigroup can be implemented easily. It allows the evaluation of hash function without storing large amount of data (look-up table) and the concept of homotopy enables consideration of many quasigropus.

Index-based n-gram extraction from large document collections

N-grams are applied in some applications searching in text documents, especially in cases when one must work with phrases, e.g. in plagiarism detection. N-gram is a sequence of n terms (or generally tokens) from a document. We get a set of n-grams by moving a floating window from the begin to the end of the document. During the extraction we must remove duplicate n-grams and we must store additional values to each n-gram type, e.g. n-gram type frequency for each document and so on, it depends on a query model used. Previous works utilize a sorting algorithm to compute the n-gram frequency. These approaches must handle a high number of the same n-grams resulting in high time and space overhead. Moreover, these techniques are often main-memory only, it means they must be executed for small or middle size collections. In this paper, we show an index-based method to the n-gram extraction for large collections. This method utilizes common data structures like B+-tree and Hash table. We show the scalability of our method by presenting experiments with the gigabytes collection.

Testing the properties of large quasigroups

With the growing importance of data security a growing effort to find new approaches to the cryptographic algorithms designs appears. One of the trends is to research the use of other algebraic structures than the traditional, such as a quasigroup. Quasigroups are equivalent to the more familiar Latin squares. There are many characteristics that must quasigroups have from the cryptography point of view. They have to be non-commutative, non-associative, non-idempotent and so on. If one want to work with quasigroups of a large order, effective methods of testing their properties are necessary. In this paper we present several experiments on various types of guasigroups and their results.

Quasigroups with good statistical properties

Quasigroups are a well-known combinatorial design equivalent to more familiar Latin squares. Since all possible elements of a quasigroup occur with equal probability, it could be an interesting tool for information assurance and network security. Most of the previous implementations of quasigroups were based on look-up table of the quasigroup, on system of distinct representatives etc. which is infeasible for large quasigroups. This paper presents statistical tests which enable us selection of good quasigroups.

Usage of MCA8 software for improving reliability of electrical networks

The MCA8 software application is described and applied to model task in this paper. This software application was developed for the purpose of the support of multi-criteria decision-making in the field of electrical power engineering. The MCA8 runs on Windows with .NET framework 2.0 and it is user-friendly. The MCA8 offers six methods of multi-criteria analysis (MCA) for solving multi-criteria decision-making tasks. These methods are WSA, IPA, TOPSIS, AGREPREF, CDA and PROMETHEE. We can use them for example for selecting the most suitable old devices in electrical networks, which we need to replace by new devices (Reclosers in this paper). The application of these remote-controlled devices causes acceleration in handling and thus shortening of duration of a fault in the network. This results in rising of probability of faultless service and thus the reliability of electrical energy supply.

Genetic search for quasigroups with heterogeneous power sequences

Genetic algorithms have been successfully used for searching for quasigroups with good properties. In this study we extend previous work done on evolutionary search for quasigroups by defining new fitness functions based on selected algebraical properties of the quasigroups. Introduced fitness functions make use of heterogeneous sequences generated during the exponentiation of quasigroups elements. Effects of the artificial evolution utilizing proposed fitness functions are illustrated on quasigroups of the order 8, 10, and 12.

Using clustering to improve WLZ77 compression

Many types of information retrieval systems (IRS) are created and more and more documents are stored in them too. The fundamental process of IRS is building of textual database, and compression of the documents stored in the database. One possibility for compression of textual data is word-based compression. Several algorithms for word-based compression algorithms based on Huffman encoding, LZW or BWT algorithm was proposed. In this paper, we describe word-based compression method based on LZ77 algorithm. IRS can also perform cluster analysis of textual database to improve quality of answers to userspsila queries. The information retrieved from the clustering can be very helpful in compression. Word-based compression using information about cluster hierarchy is presented in this paper. Experimental results which are provided at the end of the paper were achieved not only using well-known word-based compression algorithms WBW and WLZW but also using quite new WLZ77 algorithm.

Social communities detection in Enron corpus using h-Index

Development of modern technologies is expanded with communications possibilities. Electronic systems of communications make possible overcoming of traditional barriers of communication, for example, such as distance. On their basis there are new types of communities which any more have no geographical restrictions. Increasing popularity of electronic communities among which projects LiveJournal, LiveInternet, and also projects popular in Russian-speaking part Internet Mamba, MirTesen, VKontakte, Odnoklassniki, etc., makes as never earlier actual questions on working out of techniques of research of similar social networks. However communications of elements of such communities only on means of electronic communications create difficulties at definition of such communities. There is some question of the specific communities connected with identification and their properties. Nevertheless such communities can be presented graphs, i.e. objects - vertices, and communications channels between them - edges. Thus a number of problems can be solved by means of use of methods from the theory of graphs. In this paper application for such networks of the methods offered for measurement of the importance of scientists work, for the purpose of definition of popularity of objects in community is described and shown. Approach demonstration is made on Enron corpus.