Software & Systems

Journal articles №4 2013

51. Simulation exercise with operator errors control [№4 за 2013 год]
Authors: (avrora-so31@mail.ru) - , Ph.D; (avrora-so31@mail.ru) - , Ph.D; (avrora-so31@mail.ru) - ; (avrora-so31@mail.ru) - ;
Abstract: The article offers the approach to automation of operator's error control when simulation training based on a model method. In this case the model is the algorithm written based on operating instructions. Deviations of an operator from the model are his errors to be minimized as a result of training. The following types of errors are considered: missing an operation, missing an operation considering its quality, relocation and adding operations. Detection of errors in serial and branching structures is shown in the article. The branching structures are described using logic operations of comparison, functional checks operations and performance monitoring operations. Complex operation "Performance-Check" is considered as an example of brunching algorithm. Functional structure of training mode structure is presented. The issues are scrutinized considering a formation of model solution of training task and comparison of training process with model solution. Formation of model solution is performed by model solution editor that is a built-in utility of training task editor. In case the model solution is present in a current training task, it is offered to form automatic evaluation of task fulfillment by comparing actual training process with this solution. General evaluation of task fulfillment, determined by evaluation of single actions, is formed according to the worst of existing partial estimates, taking into account total limit of time allotted for fulfillment of training task.
Keywords: control management, operation, functional network, standard, model, pair- wise comparison, algorithm, mistake (error), operator
Visitors: 9626

52. Three-dimensional imaging method for the parameters of fuel elements of VVER-1000 reactors [№4 за 2013 год]
Authors: V.V. Levshchanov (master_v@inbox.ru) - Technological Research Institute of Ulyanovsk State University (Senior Researcher), Ph.D; D.S. Lavygin (andrew.tsyganov@gmail.com) - Technological Research Institute of Ulyanovsk State University (Senior Researcher), Ph.D;
Abstract: The article considers the issues of 3D graphical representation of operational and geometric characteristics of fuel elements and fuel assemblies of VVER-1000, which are implemented in Fuel Elements Data System software. The paper proposes modern approaches to writing an application with a graphical user interface, which implements the Firebird database client, based on open software libraries such as Nokia Qt, QwtPlot and GLC_lib. Thus, using these tools allows creating a number of new features in charts constructing, including: easy chart scaling and automatic setting the limits on the coordinate axes. The main tool in the design of the visualizer window interface was GLC_lib software library. It provided features for the displaying assembly of fuel elements, including the functions to control the position of the three-dimensional scene camera, lighting, texturing and color. Any object in the scene can be selected by a cursor. This software library within Nokia Qt SDK which includes Nokia Qt Creator software environment, provides additional functional features for developing software product.Special attention is paid to problems of representation information extracted from the database.
Keywords: qwtplot, glc_lib, qt, opengl, three-dimensional scenes, software libraries, fuel elements, visualizer
Visitors: 7156

53. Preprocessing the video sequence to publicate in digital libraries [№4 за 2013 год]
Authors: Damov M.V. (me@damov.pro) - Siberian State Aerospace University named after academician M.F. Reshetnev (Assistant Professor), Ph.D; V.V. Buryachenko (buryachenko@sibsau.ru) - Academician M.F. Reshetnev Siberian State University of Science and Technology (Associate Professor), Ph.D; Pakhirka A.I. (pahirka@sibsau.ru) - Siberian State Aerospace University named after academician M.F. Reshetnev (Assistant Professor), Ph.D; Zotin A.G. (zotinkrs@gmail.com) - Academician M.F. Reshetnev Siberian State Aerospace University (Associate Professor), Ph.D;
Abstract: This paper presents methods for increasing the quality of embedded video sequences when preparing digital books and other digital library multimedia resources. They are presented as processing stages: non-linear correction of image brightness, scene replacement detection with preliminary motion estimation in the scene and preliminary quality estimation of the frame, image stabilization with frame borders restoration methods. Non-linear correction of images brightness performs using modified MSR-algorithm which allows increasing the contrast of image details in of high brightness areas. The algorithm of video sequence stabilization is based on block matching method and allows removing frame-to-frame jitter effect. An interpolation of neighboring frames is used when stabilizing. It provides less image content loss on frame borders. It is offered to evaluate the quality of video sequence frame that allows detecting cases of scene changes and high level noise. A description of testing and results of the increase of video sequences quality it presented.
Keywords: digital libraries, digital publication, motion correction, image brightness correction, frame interpolation, motion evaluation, video stabilization
Visitors: 11503

54. formal methods of text authorship attribution and their usage in software products [№4 за 2013 год]
Authors: Batura T.V. (tatiana.v.batura@gmail.com) - A.P. Ershov Institute of Informatics Systems (IIS), Siberian Branch of the Russian Federationn Academy of Sciences, Ph.D;
Abstract: The article represents the survey of authorship attribution methods. It also provides a description of the popular software systems to determine the author's style, focused on the Russian language. There was an attempt to make their comparative analysis, to identify advantages and drawbacks of approaches. An authorship attribution is often based on the assumption that the alleged author is known. The analysis of syntactic, lexical-phraseological and stylistic levels of text is the most interesting and the most difficult. Expert analysis of the author's style is a time consuming process, so the attention is paid to the formal methods of attribution. Currently, to establishing the authorship of texts the following methods are used: the approaches of pattern recognition theory, methods of mathematical statistics and probability theory, neural network algorithms, cluster analysis algorithms, etc. Among the challenges to research the field of authorship attribution there is a problem of choosing formal parameters characterizing the text and style of the author, the sampling problem of reference texts. The software products are not focused on integrated study and results comparison. They are applied to various problems of analysis of text styles with different frequency characteristics and different test material. To find a new or improve existing methods of attribution of texts further research is needed. It is also important to carry out experiments aimed at finding features that allow to clearly separate the authors styles, including the small sample size.
Keywords: characteristics of text, text categorization, author's style, formal parameters of the text, authorship attribution, text attribution
Visitors: 24563

55. Determination of sentiment words weights based on genetic algorithm in the sentiment analysis [№4 за 2013 год]
Authors: (Kotelnikov.EV@gmail.com) - , Ph.D; (klekovkina.mv@gmail.com) - ;
Abstract: The article is dedicated to the problem of a text sentiment analysis using a dictionary method. The sentiment analysis is used in many areas, for example, in the marketing research, the recommendatory systems, the search engines, the human-computer interaction, the news analysis, etc. The dictionary method along with machine learning is one of the most effective approaches to the sentiment analysis. The dictionary method is based on the sentiment lexicon; each word from this lexicon has a weight indicating the degree of the word importance. The quality of text sentiment analysis depends significantly on the correct assignment of the words’ weights. This paper proposes a new method for assignment of the words’ weights. The problem of the weight assignment is considered as a multi-dimensional function optimization task. The arguments of the function are the weights of words; the values of the function are the some metric of the quality of sentiment analysis. This metric is obtained from dictionary method with given weights. In this paper the metric is F1‑measure. The optimization problem is solved using a genetic algorithm. The experiments were carried out on a collection of movie reviews from the seminar ROMIP‑2011. The results demonstrate the advantage of the proposed method of weight assignment over other ways, as well as the superiority of the dictionary method over the support vector machine.
Keywords: support vector machines, generic algorithm, dictionary method, text sentiment analysis
Visitors: 11231

56. Automatic continuous speech recognition system for the tatar language [№4 за 2013 год]
Authors: (khusainov.aidar@gmail.com) - ; D.Sh. Sulejmanov (khusainov.aidar@gmail.com) - Institute of Applied Semiotics of the Tatarstan Academy of Sciences, Kazan Federal University (Professor), Ph.D;
Abstract: The paper describes speech analysis system for the Tatar language. An approach to creating automatic phonetic transcription system for Tatar texts has been developed and implemented. This system is based on acoustic rules for the Tatar language. These acoustic rules have been converted to a formalized form using developed programming tool. The two levels analysis procedure has also been developed in order to create speech recognition algorithm, these levels are: acoustic (phonemes) level, linguistic (words) level. Phonemes level consists of acoustic models for 57 Tatar phonemes, each of this phonemes is represented by a hidden Markov model. To train each of specified statistical model 5 hours speech corpus has been recorded. Additionally, an algorithm to automatically create phoneme-level annotation of this corpus has been developed and applied. Finally, a Tatar speech recognition algorithm has been implemented based on developed programming tools and acoustic models. In total, the proposed system has shown 92% word recognition correctness in test speech subcorpus.
Keywords: hidden Markov models, speech analysis, phonetic transcription, Tatar language, speech corpus, speech recognition
Visitors: 10362

57. Typologization of local maximums in electroencephalogram spectrums derived from continuous wavelet transformation [№4 за 2013 год]
Authors: Turovskiy Ya.A. (yaroslav_turovsk@mail.ru) - Voronezh State University, Ph.D; Kurgalin S.D. (kurgalin@bk.ru) - Voronezh State University, Ph.D; (revolusion@mail.ru) - ;
Abstract: The paper offers a method to obtain a functionally important information about brain work. It is based on the analysis of local maxima chains (LMC) on the set of the wavelet energy spectrum (scalogramm, local spectra) derived from a matrix of squares of the coefficients of the wavelet transformation of electroencephalogram (EEG) data.The method means averaging LMC in frequency space. The method uses two ways of averaging: by the first LMC timing coordinate, and by minimizing the differences between LMC. When using the first way, all chains are synchronized according to the first value which equals to zero scaling reading the timeline; in the second way chains move relatively to each other in order to minimize the differences between them, and only then averaging is run. The research have shown that for the EEG of a human at rest there are five basic types of LMC behavior: 1) the increasing one, characterized by increasing frequency of the local spectrum maxima in the chain, 2) the decreasing one, characterized by reducing frequency of local maxima in time, 3) the constant one, when the whole LMC frequency is not changing significantly, 4) ascending-descending, when the local maxima frequency in the chain first increases and then decreases, 5) the decreasing-increasing one which is opposite to the behavior of the 4th type. It is determined that the relation of the chain to a certain type is an individual property of a human and depends on the nature of the brain activity at the moment. The typology performed for the local maxima of the matrices of the wavelet coefficients for EEG is essential to study the EEG in different functional states of a human.
Keywords: scalegram, electroencephalogram, medical and biological information, wavelet-analysis
Visitors: 7416

58. The interpreter of descriptions of expert systems for psychological and physiological diagnostics [№4 за 2013 год]
Authors: (alerma@rambler.ru) - , Ph.D; (ksennaidd@gmail.com) - , Ph.D;
Abstract: The article describes the developed problem-oriented interpreter of expert systems descriptions, which is an integral part of the automated technology of creating expert systems for psychological and physiological diagnostics, providing practical use of these systems: maintaining an inspects database, diagnostic procedures implementation, on-screen display requests to the user with diagnostic tools and information messages, step-by-step data processing, control of the survey course, diagnostic conclusion output, the displaying survey results on the monitor screen and printing device in tabular and graphical form as well as their storage and view in the database. The descriptions interpreter also allows implementing diagnostic-training systems for trainings of students and young professionals. Functioning of the descriptions interpreter is based on the proposed formalized model of solved diagnostic tasks and the complex of diagnostic conclusion output models. The article describes the automated technology of creating expert systems of psychological and physiological diagnostics, characterizes this class of expert systems, describes the computational procedures and operations supported by the descriptions interpreter, the varieties of used diagnostic conclusions inference rules. The article also describes the concept of creating the descriptions interpreter, its basic functionalities and the structure at the functional subsystems level. The article discusses the software implementation features of the descriptions interpreter developed in the Microsoft Visual Basic 6.0 Service Pack 6 programming language and intended for working in Windows XP operating system.
Keywords: psycho-physiological diagnostics, knowledge base, model of inference of diagnostic conclusion, formalized model of knowledge, the automated creation of expert systems, expert system, the interpreter descriptions of expert systems
Visitors: 12079

◄ ← Preview | 1 | 2 | 3 | 4 | 5 | 6