Accuracy of the Parsing of Lithuanian Simple Sentences

Abstract

The problem of the parsing accuracy of simple sentences is solved. The case of the language with high inflexion is investigated. The task is addressed by the example of the Lithuanian language, which one root can give hundreds or even more than one thousand word forms. The method of estimation of the parsing accuracy of the simple sentences of such language is given, which is based on the usage of knowledge of language consistent patterns. It is taken note, that in the case of language with high inflection and small number of users the usage of the statistical data on language is strongly restricted. The method of the accuracy estimation of the parsing of simple sentences is presented. The algorithm of implementation of the software is described. The validity of the propositions is proved by experiments. The material of the Lithuanian corpus was used for the experiments. The recommendations are given for the increasing of the accuracy of the parsing of simple sentences for the languages with high inflection.

DOI: http://dx.doi.org/10.5755/j01.itc.43.4.6700