The algorithm for selection of n-words and formation of a description of the subject area in the form of a fuzzy ontology based on the hybridization of linguistic and statistical methods of text analysis has been developed. The input data of the algorithm are text sequences in machine language, the result of the algorithm operation is the description of the subject area in the form of a fuzzy ontology. The definition of classes of fuzzy ontologies in accordance with the source code is carried out using several machine learning algorithms that combine methods of data preparation and prediction. Methods “Bootstrap”, “Bagging” and “Random forest” were used to classify text sequences. A feature of the developed algorithm is the need to represent ontology objects mainly in the form of one-words with maximization of the number of relations between objects.
The article presents the results of analyzing the process of detecting patterns in semi-structured and unstructured texts. Text preprocessing is aimed at removing terms that do not carry useful information, bringing the words of the text to a single representation and forming a set of indices of a text information resource - statistical, linguistic, structural. An algorithm for graphematic analysis has been developed as the first stage of automatic processing of texts in natural language. The algorithm makes it possible to single out semantically significant natural language constructions in a semi-structured text using graphematic descriptors, to detect and replace abbreviations and abbreviations.
The article describes the tasks of the optical recognition system for tabular documents and the development of a formal model of the optical recognition mechanism for tabular documents. The aim of the study is to use the developed system of automatic analysis of platform-type documentation as expert and training to support decision making, semantic analysis or audit of any content, including in other industries based on the principle of automatic filling of ontologies. The proposed system will allow to overcome this barrier in a cardinal way – by automatically creating a database. In this case, the user’s role will be reduced to the choice of sources for filling ontologies (the choice of thesauruses, corpus of texts in machine-readable or natural language, etc.) and the configuration of inference rules, i.e., essentially, to the configuration operations.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.