Natural Language Processing for the Semantic Web. Diana Maynard

morphological analysis, in addition to the main tools required for an information extraction system (named entity recognition and relation extraction) which build on these components. The second half of the book explains how Semantic Web and NLP technologies can enhance each other, for example via semantic annotation, ontology linking, and population. These chapters also discuss sentiment analysis, a key component in making sense of textual data, and the difficulties of performing NLP on social media, as well as some proposed solutions. The book finishes by investigating some applications of these tools, focusing on semantic search and visualization, modeling user behavior, and an outlook on the future.

       KEYWORDS

      natural language processing, semantic web, semantic search, social media analysis, text mining, linked data, entity linking, information extraction, sentiment analysis

      This book is a timely exposition of natural language processing and its role and importance for those seeking to apply semantic technologies. Clearly written with good coverage of the key topics and a comprehensive bibliography, the text will be invaluable for semantic web practitioners and more widely.

       Prof John Davies

       BT Research & Technology

       Adastral Park UK

       November 2016

       Contents

       Acknowledgments

       1 Introduction

       1.1 Information Extraction

       1.2 Ambiguity

       1.3 Performance

       1.4 Structure of the Book

       2 Linguistic Processing

       2.1 Introduction

       2.2 Approaches to Linguistic Processing

       2.3 NLP Pipelines

       2.4 Tokenization

       2.5 Sentence Splitting

       2.6 POS Tagging

       2.7 Morphological Analysis and Stemming

       2.7.1 Stemming

       2.8 Syntactic Parsing

       2.9 Chunking

       2.10 Summary

       3 Named Entity Recognition and Classification

       3.1 Introduction

       3.2 Types of Named Entities

       3.3 Named Entity Evaluations and Corpora

       3.4 Challenges in NERC

       3.5 Related Tasks

       3.6 Approaches to NERC

       3.6.1 Rule-based Approaches to NERC

       3.6.2 Supervised Learning Methods for NERC

       3.7 Tools for NERC

       3.8 NERC on Social Media

       3.9 Performance

       3.10 Summary

       4 Relation Extraction

       4.1 Introduction

       4.2 Relation Extraction Pipeline

       4.3 Relationship between Relation Extraction and other IE Tasks

       4.4 The Role of Knowledge Bases in Relation Extraction

       4.5 Relation Schemas

       4.6 Relation Extraction Methods

       4.6.1 Bootstrapping Approaches

       4.7 Rule-based Approaches

       4.8 Supervised Approaches

       4.9 Unsupervised Approaches

       4.10 Distant Supervision Approaches

       4.10.1 Universal Schemas

       4.10.2 Hybrid Approaches

       4.11 Performance

       4.12 Summary

       5 Entity Linking

       5.1 Named Entity Linking and Semantic Linking

       5.2 NEL Datasets

       5.3 LOD-based Approaches

       5.3.1 DBpedia Spotlight

       5.3.2 YODIE: A LOD-based Entity Disambiguation Framework

       5.3.3 Other Key LOD-based Approaches

       5.4 Commercial Entity Linking Services

       5.5 NEL for Social Media Content

       5.6 Discussion

       6 Automated