Research of unstructured data interpretation problems

Abstract

The term «unstructured data» means data that is unordered and arbitrary in shape. However, this type of information has a certain structure. Today there is a wide variety of data and, as a result, it is necessary to interpret them. Interpretation tasks include forecasting, classification, clustering, association, sequence search, data visualization, and variance analysis. The difficulty lies in the fact that the data itself can differ not only in terms of format, but also in terms of its structure. One of the key tasks when working with unstructured data is to find and identify patterns in order to understand them and develop filling patterns. The paper analyzes the rules for the design of bibliographic sources in order to identify common patterns. The concepts of structured and unstructured data are touched upon. The existing directions of work with unstructured data and methods of processing unstructured data, in particular, the rules for the design of bibliographic lists of literary sources, are considered. These rules were used to form templates consisting of semantic groups on the basis of examples of the corresponding lists of bibliographic sources. The final comparison of the obtained templates revealed both common features that unite all the considered templates and features that separate them.