Semantic Analysis and Text Summarization in Socio-Technical Systems
In this chapter the authors present the results of the development the methodology for increasing the reliability of the functioning of the Socio-Technical System. The existed methods and algorithms for processing unstructured (textual) information were studied. Taking into account noted above strengths and weaknesses of Discriminant and Probabilistic approaches of Latent Semantic Relations analysis in of the summarization projection the Methodology of Two-level Text Summarization Based on the Latent Semantic Relations was developed. During developing and testing this methodology authors proved, that the combination of the discriminant and probabilistic methods gives the opportunity to improve the text summarization quality by using the synergistic effect of semantic approaches and latent relations identification as well as algorithms for prediction of random processes. This effect was achieved via increasing: the quality of grouping (clustering) of fragments of the original analyzed text (paragraphs or sentences) by using a complex application of LDA-method recognizing taking account the hidden latent semantic relations phenomena; the quality of text summarization process by using a complex application latent semantic analysis taking account the optimal number of probabilistic topics. The examples of the proposed Methodology realization were presented. Formation of this kind of summary could help to human as an element of Socio-Technical System to increase the speed of processing the big amount of information (for example, realization the procedure of monitoring and analysis the daily or retrospective documents, as well as supporting the making-decision processes).
wyświetlono 5 razy