On modeling multilingual intelligent content analysis

This article presents an experience of developing a model for intelligent content analysis, which is a necessary resource for computer technologies of processing unstructured information. A distinctive feature of the model is the possibility of its application for the analysis of texts in various national languages and the mechanism for extracting task-oriented content that is not explicitly presented at the surface level of the analyzed text. The model consists of two main components: first, a knowledge base that includes a domain-oriented multilingual ontology, ontolexicons, dynamic frames, and, second, rules for text processing and presentation of content analysis results. The methodology for developing a multilingual model and the actual procedure of content analysis are described in regard to their application to the corpus of news reports on terrorist attacks in English.