Content
Summary i
Samenvatting iii
Acknowledgments v
Contents vii
1 Introduction 1
1.1 Background and Motivation 1
1.2 Research Objectives 4
1.3 Research approach 5
1.4 Research relevance 6
1.5 Thesis Outline 6
2 Research corpus, its study and tools used 9
2.1 Introduction 9
2.2 Structure of gazetteer entries 11
2.3 Digitisation and storage of Gazetteer data 12
2.4 Spatial hints and their components 13
2.5 Dataset used 17
2.6 Tools used 18
2.7 Summary 18
3 Extracting spatial hint information from locality descriptions 21
3.1 Introduction 21
3.2 Related work 22
3.3 Approach and implementation 24
3.4 Description normalisation 26
3.5 Natural Language Processing pipeline 28
3.6 Identifying subtree patterns in parse trees 32
3.7 Information Extraction pipeline 41
3.8 Evaluation and Results 45
3.9 Summary 47
4 Spatial extent models for directional containment phrases 49
4.1 Introduction 49
4.2 Related work 50
4.3 Background for constructing spatial extent models 52
4.4 Construction of spatial extent models 55
4.5 Model performance and comparison 62
4.6 Data analyses 69
4.7 Summary 78
5 Spatial representation of hints using probability rasters 81
5.1 Introduction 81
5.2 Background on uncertainty 82
5.3 Related work 86
5.4 Uncertainty in spatial hints 88
5.5 General approach to probability raster creation 90
5.6 Raster creation for spatial hints 92
5.7 Optimisation of input parameters 101
5.8 Raster overlay 105
5.9 Overlaying rasters with vector model 111
5.10 Summary 113
6 Synthesis 115
6.1 Conclusions 115
6.2 Limitations 118
6.3 Recommendations for future work 119
Bibliography 121
A Appendix 131
B Appendix 135
C Appendix 137