The article describes peculiarities of modern syntax parser systems and problems originating in text analysis. As a result of comparative analysis the authors propose a unified approach to processing of unstructured texts in Russian and English which combines morphology and syntax processing. The developed syntax analysis system, using verbs’ valency dictionary, samples of minimal structural schemes of sentences and samples of conjunctions, allows choosing predicative structures of sentences in the text, realizing initial semantic analysis due to semantic content of predicate’s actants and building trees of syntactical subordination of sentences. The derived trees hold elements of tree of constitutives and tree of dependences. The proposed samples and rules organization allows resolving some of the problems of modern parsers. And the use of verbs’ valency dictionary allows reducing the number of sentences syntax analysis variants.

Keywords: automatic text processing; syntax parser; morphological analysis; structural text elements

The represented approach of dynamic process modeling is based on the technology of automatical semantic text analysis. An associative network is forming during text processing. Its key notions, including  lexical and psycholinguistic markers of the analyzed process, are ranked by theirs semantic weight. The weight being multiplied by marker status value at the scale of “good-bad” gives its contribution to the process stage characteristic. Transformation of the accumulated for all of the markers process characteristic from one period of time to another one is characterize a direction of the process.

Keywords: automatical text processing, associative (homogenous semantic) network, process dynamic modeling, social processes, lexical and psycholinguistic markers

Common conceptions are given about multi-sensor systems. The structure of the data processing unit of multi-sensor system for concentration monitoring of heavy metals ions in aqueous media is described. Main stages of the data processing, – clustering and quantitative identification are introduced. The approach which presupposes use of artificial neural networks, – Kohonen networks and radial basis function networks, is designed. Such algorithmic architecture is implemented which allows to avoid significant computing resources expenses. The program for computer IBM PC in high-level language of application package MATLAB is written.

Keywords: Multi-sensor systems, artificial neural networks, Kohonen networks, radial basis function networks, data processing, pattern recognition, environmental monitoring, heavy metals, ion-selective electrodes

Information about authors of issue №4 (2013)

Keywords: authors

Information about authors of issue №4 (2011)

Keywords: authors