This work presents neural network analysis to the social scientist without a background in computer science. Studies and examples illustrate the advantages of neural network analysis over other procedures in use amongst social scientists. Other features include: an introduction to the vocabulary and framework of neural networks; a history of neural network methods; a review of the literature; examples of neural network applications in the social sciences; coverage of alternative neural network models; and examples using two software packages for neural network analysis.
G. David Garson is a full professor of public administration at North Carolina State University, where he teaches courses on advanced research methodology, geographic information systems, information technology, e-government, and American government. In 1995 he was recipient of the Donald Campbell Award from the Policy Studies Organization, American Political Science Association, for outstanding contributions to policy research methodology and in 1997 of the Aaron Wildavsky Book Award from the same organization. In 1999 he won the Okidata Instructional Web Award from the Computers and Multimedia Section of the American Political Science Association, in 2002 received an NCSU Award for Innovative Excellence in Teaching and Learning with Technology, and in 2003 received an award "For Outstanding Teaching in Political Science" from the American Political Science Association and the National Political Science Honor Society, Pi Sigma Alpha. In 2008 the NCSU Public Administration Program was named in the top 10 PA schools in the nation in information systems management. Prof. Garson is editor of and contributor to Handbook of Public Information Systems, Third Edition.(2010); Handbook of Research on Public Information Technology (2008), Patriotic Information Systems: Privacy, Access, and Security Issues of Bush Information Policy (2008), Modern Public Information Technology Systems (2007), and author of Public Information Technology and E-Governance: Managing the Virtual State (2006), editor of Public Information Systems: Policy and Management Issues (2003), coeditor of Digital Government: Principles and Practices (2003), coauthor of Crime Mapping (2003), author of Guide to Writing Quantitative Papers, Theses, and Dissertations (Dekker, 2001), editor of Social Dimensions of Information Technology (2000), Information Technology and Computer Applications in Public Administration: Issues and Trends (1999) and is author of Neural Network Analysis for Social Scientists (1998), Computer Technology and Social Issues (1995), Geographic Databases and Analytic Mapping (1992), and is author, coauthor, editor, or coeditor of 17 other books and author or coauthor of over 50 articles. He has also created award-winning American Government computer simulations, CD-ROMs, and six web sites for Prentice-Hall and Simon & Schuster (1995-1999). For the last 28 years he has also served as editor of the Social Science Computer Review and is on the editorial board of four additional journals. His widely-cited online textbook, Statnotes: Topics in Multivariate Analysis (2006-2009), is used by over 1.5 million people a year. Professor Garson received his undergraduate degree in political science from Princeton University (1965) and his doctoral degree in government from Harvard University (1969).
Introduction to Neural Network Analysis The Terminology of Neural Network Analysis The Backpropagation Model Alternative Network Paradigms Methodological Considerations Neural Network Software Example Analysing Census Data with Neural Connection Conclusion
The author is the recipient of the Donald C. Campbell Award for Methodological Innovation and the Aaron Wildavsky Book Award from the Policy Studies Organization/American Political Science Association.
`Garson's book would be a good buy for someone setting out to apply neural networks to their data. It takes a balanced approach, trying to make it clear where they would be applicable and where traditional statisitcs might be a better bet. It is certainly easy to read' - British Journal of Mathematical and Statisistical Psychology `A useful reference source for terminology, mathematical background, possible application areas and pointers towards software use' - Statistical Methods in Medical Research The much of the material within is timeless and the quality of its presentation allows it to remain a value-add contributor, even today. Overall this book needs to be taken off the storage shelf, dusted off, and placed on your lap. The book's publication age is an advantage in this case as the all-important basics of neural networks are not skimmed over in this book as they often can be the books published today. This is a must-read for any computational modeler looking to a way to progress their technique. -- Terrill L. Frantz
