Text as Data

Combine the best of qualitative and quantitative techniques within the SAS system for superior results Text analytics has become an indispensable part of fields as diverse as pharmaceutical research and development and social media marketing. Organizations around the world are implementing comprehensive, text-inclusive analytics strategies. In Text as Data: Computational Methods of Understanding Written Expression Using SAS, you’ll discover how and why the SAS platform delivers exceptional text analytics results by combining linguistic and quantitative algorithms and treating text as qualitative data from within a quantitative analytics framework. The accomplished authors offer a thorough introduction to the principles and dynamics of text analytics, along with a comprehensive overview of an effective framework for common use cases. Readers will learn about the interplay between qualitative-linguistic and quantitative data analysis and gain a deep understanding of techniques like expression parsing and detection, text parsing, theme and topic detection, and more. They’ll also discover why SAS is the ideal platform for deploying a text analytics solution. Ideal for SAS users and data scientists in any industry, Text as Data provides readers with a rich and insightful exploration of text analytics with SAS, creating a foundation for practical and effective applications. Text As Data: Combining qualitative and quantitative algorithms within the SAS system for accurate, effective and understandable text analytics The need for powerful, accurate and increasingly automatic text analysis software in modern information technology has dramatically increased. Fields as diverse as financial management, fraud and cybercrime prevention, Pharmaceutical R&D, social media marketing, customer care, and health services are implementing more comprehensive text-inclusive, analytics strategies. Text as Data: Computational Methods of Understanding Written Expression Using SAS presents an overview of text analytics and the critical role SAS software plays in combining linguistic and quantitative algorithms in the evolution of this dynamic field. Drawing on over two decades of experience in text analytics, authors Barry deVille and Gurpreet Singh Bawa examine the evolution of text mining and cloud-based solutions, and the development of SAS Visual Text Analytics. By integrating quantitative data and textual analysis with advanced computer learning principles, the authors demonstrate the combined advantages of SAS compared to standard approaches, and show how approaching text as qualitative data within a quantitative analytics framework produces more detailed, accurate, and explanatory results. * Understand the role of linguistics, machine learning, and multiple data sources in the text analytics workflow * Understand how a range of quantitative algorithms and data representations reflect contextual effects to shape meaning and understanding * Access online data and code repositories, videos, tutorials, and case studies * Learn how SAS extends quantitative algorithms to produce expanded text analytics capabilities * Redefine text in terms of data for more accurate analysis This book offers a thorough introduction to the framework and dynamics of text analytics—and the underlying principles at work—and provides an in-depth examination of the interplay between qualitative-linguistic and quantitative, data-driven aspects of data analysis. The treatment begins with a discussion on expression parsing and detection and provides insight into the core principles and practices of text parsing, theme, and topic detection. It includes advanced topics such as contextual effects in numeric and textual data manipulation, fine-tuning text meaning and disambiguation. As the first resource to leverage the power of SAS for text analytics, Text as Data is an essential resource for SAS users and data scientists in any industry or academic application.  BARRY DEVILLE is a Data Scientist and Solutions Architect with 18 years of experience working at SAS. He led the development of the KnowledgeSEEKER decision tree package and has given workshops and tutorials on decision trees for Statistics Canada, the American Marketing Association, the IEEE, and the Direct Marketing Association. GURPREET SINGH BAWA is the Data Science Senior Manager at Accenture PLC in India. He delivers advanced analytics solutions for global clients in a variety of corporate sectors. Preface xi Acknowledgments xiii About the Authors xv Introduction 1 Chapter 1 Text Mining and Text Analytics 3 Chapter 2 Text Analytics Process Overview 15 Chapter 3 Text Data Source Capture 33 Chapter 4 Document Content and Characterization 43 Chapter 5 Textual Abstraction: Latent Structure, Dimension Reduction 73 Chapter 6 Classification and Prediction 103 Chapter 7 Boolean Methods of Classification and Prediction 125 Chapter 8 Speech to Text 139 Appendix A Mood State Identification in Text 157 Appendix B A Design Approach to Characterizing Users Based on Audio Interactions on a Conversational AI Platform 175 Appendix C SAS Patents in Text Analytics 189 Glossary 197 Index 203