The Formation of Problem Domain Dictionary
Наталья Владимировна Борисова, Olga Kanishcheva · 2013
The currently available methods for automatic generation of terms and descriptions of the compatibility of semantic classes do result in the sufficient quality to completely eliminate manual work of experts. They however can give a first approximation to the required specification and to provide an expert with statistical data for further improvement of this description.Thus, the development of methods for the automated formation of compatibility terms descriptions is timely, allowing to extract information about the compatibility of text collections, summarize and present it the way that experts can effectively work with. Another problem is taking into account the already formed and also the combinability descriptions existing in dictionaries to improve the effectiveness of automatic parsing.This paper presents an approach to the formation of a compatibility terms dictionary of the subject area. The basis of the approach is the mathematical formalism of the algebra of finite predicates and predicate operations.The algebra of predicates and predicate operations allows the to description of the semantic combinability of the two words standing next to each other which is determined by the composition of semantic roles. Between sets of semantic roles of words standing next to each other there is a binary relation which is a subset of the Cartesian product of these sets. The presence or absence of coordination between words is calculated using the comparator identification method.