1. ABSTRACT
ACKNOWLEDGEMENTS
DECLARATION
CONTENT
LIST OF TABLES
LIST OF FIGURES
PUBLICATION
ABBREVIATIONS
NOTATIONS
CHAPTER 1 INTRODUCTION
1.1 Chemometrics and Pattern Recognition
1.1.1 Exploratory data analysis
1.1.2 Multivariate regressions
1.1.3 Classification
1.1.4 Variable selection
1.2 Environmental Studies
1.3 Quantitative Structure-Activity Relationship studies
1.4 Aims of this thesis
1.4.1 Data pre-processing and data drift: The utility of Principal Component Analysis, Self
Organizing Maps and class separation indices
1.4.2 Multiblock methods and regression analysis of environmental dataset
1.4.3 Pattern recognition in QSAR with an application of SOMs and PLSDA
1.4.4 Data splitting methods for regression and classification 18
CHAPTER 2 DATASET
2.1 Introduction
2.2 Environmental studies
2.2.1 Airborne particulate matter datasets
2.2.2 Meteorological dataset
2.3 Quantitative Structure Activity Relationships studies
2.3.1 Biological activity data collection
2.3.2 Software and tools
2.3.4 Structural conversion and data curation
CHAPTER 3 PATTERN RECOGNITION AND DATA MINING METHODS IN
CHEMOMETRICS
3.1 Introduction
3.2 Data pre-processing
3.2.1 Individual transformation
3.2.2 Row scaling
3.2.3 Column scaling
3.2.3.1 Mean centring
3.2.3.2 Standardisation
3.3 Exploratory data analysis
3.3.1 Principal Component Analysis
3.3.2 Self Organizing Maps
CHAPTER 4 DATA PRE-PROCESSING AND DATA DRIFT: EXPLORATORY
ANALYSIS AND CLASS SEPARATION INDICES
4.1 Introduction
4.2 Dataset
4.2.1 Airborne particulate matter dataset 76
4.2.2 Meteorology dataset 78
2. 4.2.3 QSAR dataset 78
CHAPTER 8 CONCLUSIONS AND FUTURE WORK 198
REFERENCES
Appendix