The unstructured texts' semantic models for the R&Ds' evaluation

The development of technological application of semantic models of text information sources for analytical and informational basement of R&D's concrete results evaluation is the application of text collection semantic models as an instrument for the expertise of R&D's persons. We provide the corresponding problem setting. For experimental justification of the suggested approach we provide text collection semantic models given in electronic form in expertise system of the Directorate of State Scientific and Technical Programs. On the base of computational models of semantics of individual documents we computationally form another models. The last models allow to split the documents into the thematically homogeneous groups as a result of completely computational procedure. Also they provide the objective view onto the directions of works financed by the Russian Ministry of Education and Science within the frame of Federal purpose-oriented program and within the frames of specific priority directions in science and technology. Given examples approve the possibility to apply the models in the questions of informational and analytical ensuring of R&D evaluation.