The Method of a Two-Level Text-Meaning Similarity Approximation of the Customers’ Opinions

The method of two-level text-meaning similarity approximation, consisting in the implementation of the classification of the stages of text opinions of customers and identifying their rank quality level was developed. Proposed and proved the significance of major hypotheses, put as the basis of the developed methodology, notably about the significance of suggestions about the existence of analogies between mathematical bases of the theory of Latent Semantic Analysis, based on the analysis of semantic relationship between the variables and degree of participation of the document or term in the corresponding concept of the document data, and instruments of the theory of Social Network Analysis, directed at revealing the features of objects on the basis of information about structure and strength of their interaction. The Contextual Cluster Structure, as well as Quantitative Ranking evaluation for interpreting the quality level of estimated customers' opinion has formed