SEARCH INDEX CONSTRUCTION METHOD AND SYSTEM USING MACHINE LEARNING ALGORITHM

摘要

FIELD: calculating; counting.;SUBSTANCE: invention relates to computer engineering. Technical result is achieved by obtaining a plurality of digital documents, wherein each corresponding document has a corresponding size, determining a corresponding significance parameter based on user interactions, ranking a plurality of digital documents to obtain a ranked list of documents based on: corresponding qualitative significance parameter based on corresponding significance parameter and corresponding size, wherein ranging is performed so that qualitative parameter of total significance is maximized on the basis of: weighting of the qualitative parameter of significance of a certain digital document based on the sum of the dimensions of a certain subset of documents, wherein said subset consists of documents having a corresponding significance parameter lower than a significance parameter of said digital document, selecting a subset of documents from the ranked list and storing said subset in the search index.;EFFECT: technical result consists in improvement of accuracy of ranging documents in a search system.;24 cl, 8 dwg