Software Journal:
Theory and Applications

Подать статью

Вход Регистрация

Outliers detection by voting method during hierarchical data clustering

A.A. Rybakov Joint Supercomputer Center of the Russian Academy of Sciences – branch of Federal State Institution «Scientific Research Institute for System Analysis of the Russian Academy of Sciences», Moscow, Россия, физико-математических наук;
S.S. Shumilin Joint Supercomputer Center of the Russian Academy of Sciences – branch of Federal State Institution «Scientific Research Institute for System Analysis of the Russian Academy of Sciences», Moscow, Россия;

At present, we often face the task of extracting useful information from a large amount of raw data. This process, called Data Mining, combines various approaches to the analysis and processing of data, but it always begins with one specific step – data cleansing. The raw data entering the analysis are often incomplete, weakly structured, they contain duplicate information and anomalies. The presence of anomalies in the array of input data can lead to incorrect interpretation of the extracted information, errors in prediction and greatly reduce the value of the knowledge obtained. Therefore, the development of new approaches to the elimination of anomalies, or outliers, is an actual task. This article discusses an approach to detecting outliers, based on hierarchical data clustering and using a voting method to identify the most likely candidates for the role of outliers.


Fuzzy Measure of Reliability to Analyze Aircrafts Technical Condition in the Context of Non-Stochastic Uncertainty

Sh. Khill Moscow Aviation Institute (National Research University), Moscow, Россия, технических наук;
V.N. Reshetnikov Center for Visualization and Satellite Information Technology, SRISA RAS, Moscow, Россия, физико-математических наук;
S.B. Savilkin Center for Visualization and Satellite Information Technology, SRISA RAS, Moscow, Россия, физико-математических наук;

Increased level of aircraft complexity and toughened requirements for the efficiency of technical condition technical condition monitoring raised the topical issue of analysis automation. Separate results were obtained, for example, during the development of unified algorithms for automated analysis of telemetric information coming from the board of manned spacecrafts. The basis of algorithms was a set of control methods, in particular, the multilevel control method, condition search trees, condition matrices, etc. With their help it is possible to perform operational analysis of technical condition of the most aircraft systems; however, a significant drawback in this case is that it is impossible to quantitatively assess the reliability of conclusions regarding technical condition of the systems. Stochastic reliability measurement in aircraft control practice is used rather limitedly due to the lack of sufficient statistics, as well as resource and time constraints. In this regard the task to develop the methods for aircraft technical condition control that allow getting real-time conclusions about technical condition of the object under control with a quantitative assessment of reliability in the context of non-stochastic uncertainty, is critical at the present time. One of possible variants to resolve it is to use the fuzzy measure of reliability.

Subsystem of technical condition control is one of important components of any modern aircraft control system. In this case the term "control" refers to the process of collecting and processing information in order to predict events. In aircraft control practice the stage of measurement data processing in order to obtain conclusions about technical condition of the object under control is considered as technical condition analysis. The present paper discusses modifications of well-known methods of automated analysis of the technical condition of aircraft on-board systems that are used for various purposes, based on the use of fuzzy measure of reliability.


Methods and means of distributed storage systems implementation

A.V. Baranov Joint Supercomputer Center of the Russian Academy of Sciences – branch of Federal State Institution «Scientific Research Institute for System Analysis of the Russian Academy of Sci-ences», Moscow, Россия, технических наук;
S.A. Leshchev Joint Supercomputer Center of the Russian Academy of Sciences – branch of Federal State Institution «Scientific Research Institute for System Analysis of the Russian Academy of Sci-ences», Moscow, Россия;

One of key methods to increase the efficiency of computing resources utilization in supercomputer centers is to link them into a distributed network. For user jobs and data to be able to migrate in a supercomputer center network, there should be a distributed storage system providing a single file space for all centers.

The paper considers the existing methods and means of implementing distributed storage systems and analyzes their applicability to a supercomputer center network. It states the requirements to a distributed storage system for different aspects of its development and use: data and its metadata storage, user job life cycle support, and data security. Additionally, the paper defines possible options and techniques of implementing stated requirements into a distributed storage system of supercomputer center network.