Building Intelligent .NET Applications Agents, Data Mining, Rule-Based Systems, and Speech Processing

by Rea, Sara Morgan

• ISBN13: 9780321246264
• ISBN10: 0321246268
• Edition: 1st
• Format: Paperback
• Copyright: 2005-03-25
• Publisher: Addison-Wesley Professional

Summary

Shows developers how to enhance both new and existing .NET applications with four powerful Artificial Intelligence technologies.

Author Biography

About the Author

Sara Rea is a senior developer for Public Systems Associates, a Baton Rouge–based company that produces legislative software. She is an MCSD and MCDBA and is certified in SQL Server 2000 and VB.NET. She specializes in developing Web-based applications utilizing Microsoft technologies. Since graduating from Louisiana State University in 1993 with a degree in quantitative business analysis, she has been developing software for a variety of industries, including a not-for-profit hospital, a financial company offering mortgages, a retail company selling coffee, and an Application Service Provider.

In 1998 she wrote an article on stress-testing Web sites that was featured on the front cover of MSDN Magazine, formerly named Microsoft Interactive Developer (MIND). Since then she has written articles featured in Enterprise Development and .NET Development. To reference the articles Sara Rea has written, refer to http://www.custsolutions.net. This Web site also features her latest research efforts and is a good resource for readers who want to learn more about enhanced computing.

© Copyright Pearson Education. All rights reserved.

Table of Contents

Foreword	xiii
Preface	xv
Acknowledgments	xxi
About the Author	xxiii
Chapter 1: Introduction	1	(12)
Profile Box: Historical Highlights of AI	2	(1)
The AI Effect	3	(1)
Enhanced Computing	4	(2)
Seamless Computing	6	(1)
Why Should You Consider Enhanced Computing?	7	(1)
Why Use Visual Studio.NET?	8	(1)
How This Book is Organized	8	(2)
Profile Box: Microsoft Research (MSR)	9	(1)
What You Need	10	(3)
Chapter 2: Creating Applications That Talk	13	(32)
The Microsoft Speech Application SDK (SASDK)	14	(5)
Profile Box: SALT Forum	15	(4)
Business Benefits of Speech	19	(1)
How the Speech Engine Works	20	(1)
Installing the SASDK	21	(1)
Creating a Speech Application	22	(16)
Profile Box: ScanSoft, Inc	28	(10)
Debugging and Tuning a Speech Application	38	(3)
Setting Up a Telephony Server	41	(1)
Summary	42	(3)
Chapter 3: Telephony Applications	45	(32)
Creating a Speech Application to Solve a Business Problem	46	(2)
Profile Box: Case Study: GMAC Commercial Mortgage	47	(1)
Loading the Sales Scheduling Database	48	(3)
Building a Telephony Application	51	(20)
Profile Box: Vocal Help Desk	65	(5)
Profile Box: Voice Manager Exchange, 3.0	70	(1)
Tuning User Prompts	71	(2)
Summary	73	(4)
Chapter 4: Multimodal Speech Applications	77	(32)
Why Create Multimodal Applications?	78	(1)
Sunrise Community College	79	(2)
Building a Multimodal Application	81	(21)
Profile Box: Case Study: Landstar Systems	83	(19)
The Manifest File	102	(1)
Running the Application on Pocket PC	103	(1)
Application Tuning	103	(3)
Summary	106	(3)
Chapter 5: Data-Mining Predictions	109	(32)
Introducing Data Mining with SQL Server	110	(3)
Savings Mart	113	(15)
Profile Box: Case Study: ComputerFleet	124	(4)
Working with Mining Models	128	(11)
Summary	139	(2)
Chapter 6: Applying Data-Mining Predictions	141	(30)
Working with the Sample Application	142	(3)
Generating New Shipment Schedules	145	(12)
Profile Box: Case Study: Web Usage Mining	156	(1)
Measuring Success of the New Shipment Method	157	(9)
Closed Loop Processing	166	(3)
Summary	169	(2)
Chapter 7: An Evolving Database	171	(30)
Understanding Rule-based Systems	172	(3)
Profile Box: BizTalk Server 2004	173	(2)
Zoom Lending	175	(1)
Building a Rule-based Application	176	(22)
Profile Box: Handling E-mails with Rules	192	(6)
Summary	198	(3)
Chapter 8: Building an Agent	201	(28)
Understanding Agents	202	(4)
Building a Multiple Agent Solution	206	(21)
Profile Box: Microsoft Agent	212	(15)
Summary	227	(2)
Chapter 9: The Future of Enhanced Computing	229	(26)
The Next Development Platform	230	(7)
Microsoft Research (MSR)	237	(6)
Profile Box: The Future of Speech at Microsoft	241	(2)
What about Al?	243	(2)
Profile Box: American Association for Artificial Intelligence	244	(1)
Other Areas of AI	245	(7)
Profile Box: Sonum Technologies	249	(3)
Opportunities for Developers	252	(3)
Glossary	255	(10)
Bibliography	265	(6)
Index	271

Excerpts

Preface Preface Artificial Intelligence (AI) has been around almost as long as computers. It was first introduced in the 1950s, and many scientists soon developed unrealistic expectations surrounding it. Throughout the past fifty years there have been as many advances as disappointments. Most people do not realize that AI-based technologies are being utilized every day. This is because once a technology becomes mainstream, it is generally no longer considered to be AI-related. Take the things we associate with intelligence--speaking, writing, thinking; these are actually some of the easiest forms for AI to reproduce. The real beauty of our intelligence is that, while not easily seen, it exists deep inside every living cell that performs a specific function. No individual cell is smart enough to function alone. Instead, humans are a complicated network of cells that carry the blueprints for all other cells. AI is not an all-or-nothing situation. It achieves success through an incredible number of baby steps. The fact that we are not yet able to build a being like Data in Star Trek does not mean we cannot create some very usable applications in our professional lives. Baby steps have already been taken in the form of technological advancements, including optical character recognition (OCR), speech recognition, digital encoding of pictures and data, and smaller, faster computers. All of these are helping to pave the way for AI-based applications. I work for a company that creates software for state and city legislatures. Our group is constantly under the gun to produce new and innovative software. To be competitive with the other vendors, we have to show that our product is better. Most important, we have to deliver the product on time. Recently, we were up against a tight deadline and the pressure was starting to mount. People were getting testy and irritable, and everyone seemed to be waiting on the next guy to do their part. One of my colleagues jokingly informed the other developers that we would soon be installing the newly released Microsoft telepathy product. This could be used to read the minds of all the other developers and make sure everyone was on the same page. Now, of course, there is no Microsoft telepathy software. Even with the phenomenal progress that has taken place in neuroscience, we still do not fully understand how the brain works, much less how to read one another's minds. Still, I can envision a day, not too far off, when we will have software that can anticipate what users are thinking and act on those predictions. This book is not about reading minds. It is about utilizing a few artificial intelligence technologies to enhance the value of existing and new applications. The four AI-related areas I cover in this book are not meant to represent everything the field of artificial intelligence has to offer. In fact, they are just a sampling of the somewhat unexplored potential for business-related applications. They are Agents Data mining Rule-based systems Speech processing Who Is This Book For? I wrote this book for developers who are not satisfied with "business as usual." For developers who are always looking to improve their code and their skill sets. For developers who surf development Web sites and read technical journals, hoping to make the applications they work on better. This book is for any developer, contract or in-house, who is building enterprise-wide applications. The applications could range from those running standalone within the company to those dispersed geographically as Web applications. The companies could be of any size, large or small, but large organizations will probably find the most benefits. Readers already familiar with Microsoft products, and especially with Visual Studio .NET, will get the most out of this book. It focuses on using AI-based techniques with Microsoft products. Readers already familiar with Microsoft SQL Server will be the most comfortable, since all of the applications utilize Transact-SQL (T-SQL) inside stored procedures. What You Should Know This book is not an introduction to programming with Visual Studio .NET. Many of the applications featured in the book are Web-based. It is assumed that the reader is already familiar with designing Web-based applications using ASP.NET. Readers should be familiar with such concepts as setting up a virtual directory and executing a Web application on their localhost machine. They should also understand the difference between server-side script and client-side script. The sample applications are not intentionally complex, but they access code embedded in SQL Stored Procedures. These stored procedures contain some of the application logic in the form of T-SQL statements. The reader should have a basic understanding of T-SQL and know how to view stored procedure code using Microsoft's Enterprise Manager or Query Analyzer. To simplify common operations such as data access and exception handling, the sample applications utilize Microsoft Application Blocks. These blocks contain code that demonstrates best practices for accessing common functionality. No additional installation steps are necessary for the reader to execute the sample code. However, readers interested in learning more about the functionality available in these blocks should visit the Microsoft Patterns and Practices Web site athttp://www.microsoft.com/resources/practices/code.mspx. A Note about Security The sample applications in this book were designed to demonstrate key techniques presented in each chapter. Although limited forms of security were considered, readers should not assume that the sample applications are production-ready. Readers interested in using the sample applications as a starting point for building their own solutions should thoroughly evaluate the code for security weaknesses. The Microsoft MSDN Web site (http://msdn.microsoft.com) contains many helpful resources for writing secure code. One resource readers may want to refer to first is a MSDN Webcast titled "Writing Secure Code--Best Practices." The Webcast, which originally played on October 12, 2004, is available through the Microsoft On-Demand Webcasts Web site athttp://www.microsoft.com/seminar/events/webcasts/ondemand.mspx. Why I Wrote This Book I am the type of developer who is always pushing the envelope--looking for more. I eagerly await the emergence of new products and technologies, and I am always trying to make the next application I write better than the last one. That being said, I also have a limited amount of time every day to get everything done. I wrote this book to explore and introduce some interesting technologies and products that I think are either unknown or unexplored. They all revolve around AI-based concepts that have been around for several years. This book seeks to introduce concepts that may have been viewed as too hard in the past. The world is changing fast, and the development world even faster. Professional software developers do not have time to learn the foundation of every AI-related technology. We barely have enough time to wade through the documentation that accompanies a beta product. What we need is someone to give us an intuitive and standard way for accessing these technologies. We know that AI technologies can be useful, but we need a good way to get to them. Sample Applications This book introduces some AI-based concepts using Visual Studio