Increasing the Reliability of Information Based on the Semantic Redundancy of Texts

The problem is formulated and constructive approaches, principles and mechanisms for increasing the reliability of information in automated document management systems using semantic redundancy of text elements are developed. Mechanisms based on the use of semantic links, correlation and associative power of lexical units are proposed. A technique for determining semantic, correlation links of text elements, comparing tokens, and powers of tested subsets of texts is developed. Algorithms for checking the reliability of information based on the Levenshtein distance are proposed. A semantic network based on the use of properties of structural components of natural-language text in nodes is developed, which extracts operators for highlighting semantic relations. A model for factorizing the term space for using tools for classifying the semantic proximity of terms is developed. Tools for using the open dictionary Wiktionary are implemented. Software tools in C++ and Java in the parallel computing environment (CUDA) are developed.

Hali tarjima qilinmagan