Developing a syntactic analyser for Estonian

. The aim of the present article is to give an overview of the current state of syntactic analysis of Estonian and describe problems that were encountered in the generation of syntactic rules for the syntactic analyser of Estonian. So far only the rules based on linguistics have been used. This article is focused on the statistical methods in syntactic analysis and it describes the experiments of using corpus-based patterns in syntactical disambiguation. 1 Introduction The aim of the analyser is to give syntactic description to the text which is previously analysed and disambiguated morphologically. The analyser marks each word with a syntactic tag. The tags denote subjects, objects, main verbs, adverbials, etc. The framework that the syntactic analyser is based on is Constraint Grammar [1], [2] which is described in following section. Section 3 gives an overview of current state of Estonian Constraint Grammar, section 4 discusses the efforts to use statistical methods for syntactic a..