Ebook: Inductive Dependency Parsing

This book provides an in-depth description of the framework of inductive dependency parsing, a methodology for robust and efficient syntactic analysis of unrestricted natural language text. This methodology is based on two essential components: dependency-based syntactic representations and a data-driven approach to syntactic parsing. More precisely, it is based on a deterministic parsing algorithm in combination with inductive machine learning to predict the next parser action. The book includes a theoretical analysis of all central models and algorithms, as well as a thorough empirical evaluation of memory-based dependency parsing, using data from Swedish and English. Offering the reader a one-stop reference to dependency-based parsing of natural language, it is intended for researchers and system developers in the language technology field, and is also suited for graduate or advanced undergraduate education.