SUPPLEMENTS

Instructor Supplements

* PowerPoints: Each figure and table in the text is part of a PowerPoint presentation.
* Test Bank: Several test questions are provided for each chapter.
* Solutions: Answers to selected exercises. Answers are given for most of the end of chapter exercises.
* Lesson planner: The lesson planner contains ideas for lecture format and points for discussion. The planner also provides suggestions for using selected end of chapter exercises in a laboratory setting.

Student Supplements

* CD-ROM: includes the IDA Software and Data Packages

IDA Software Package

Experiential learning is required to develop the skills required of a data mining expert. The iDA software is designed to give students this needed hands-on experience with the data mining process. The software is used in several chapters to illustrate many important data mining concepts. Chapters 4, 5, 7, 9, 10, 11 and 13 have several end of chapter exercises designed for the iDA software.
iDA consists of a preprocessor, a report generator, and three data mining toolsæESX for supervised learning and unsupervised clustering, a neural network tool for creating supervised back propagation models and unsupervised self-organizing maps, and a production rule generator. Since iDA is an Excel add-on, the user interface is Microsoft Excel. We chose iDA because of its flexibility and ease of use.

IDA Dataset Package

Several datasets are included with the iDA software. The datasets come from three general application areas æ business, medicine and health, and science. All datasets are in MS Excel format and ready to use.
Datasets can be described along several dimensions including the number of data instances, number of attributes, amount of missing or noisy data, whether data attributes are clearly defined, whether the data is categorical, numeric or a combination of both data types, whether well-defined classes exist in the data, whether a time element is implicit in the data, whether the input attributes can differentiate between known classes contained in the data, and whether input attributes are correlated. As the above factors affect the way data mining is performed, the iDA datasets were chosen to provide variety among these dimensions. The datasets also serve several general purposes. Specifically,
~ Provide the beginning student with experimental data to experience the data mining process without requiring the student to deal with data pre-processing issues.
~ Show the wide range of problem areas and problem types appropriate for data mining solution.
~ Explain data mining outcomes.
~ Illustrate the knowledge discovery process.
~ Recognize that experimentation with several data mining techniques may be necessary to create a best model for a specific dataset.


Here is a short description of the datasets that are part of the iDA software package. The description includes a short statement about one or more characteristics of each dataset.

Business Applications

The Credit Card Promotion Dataset.
This is a hypothetical dataset containing information about credit card holders who have accepted or rejected various promotional offerings. The dataset is used to illustrate many of the data mining techniques discussed in the text.

The Credit Card Screening Dataset.
The file contains data about individuals who have applied for a credit card. The output attribute indicates whether each individual’s credit card application was accepted or rejected. The input attributes have been changed to meaningless symbols to protect confidentiality of the data.

The Deer Hunters Dataset.
The dataset holds information about deer hunters who are either willing or unwilling to spend more for their next hunting trip. Several irrelevant input attributes are present in the data.

The Stock Index Dataset.
The data is a time series representation of average weekly closing prices for the Nasdaq and the Dow Jones Industrial Average.

Medicine and Health

The Cardiology Patient Dataset.
This dataset holds medical information about two groups of individuals. One group of individuals have suffered one or more heart attacks. The second group of individuals have not experienced a heart attack. The dataset contains a nice mix of categorical and numeric attributes.
The Spine Clinic Dataset. This dataset contains medical information about individuals who have had lower back surgery. Some of these folks have returned to work while others have not. A clear definition of the mean of each attribute is not given. The dataset contains both numeric and categorical data.
Science

The Gamma Ray Burst Dataset.
The dataset contains recorded information about individual gamma-ray bursts. Gamma-ray bursts are brief gamma-ray flashes whose origins are outside our solar system. The bursts were observed by the Burst And Transient Source Experiment (BATSE) aboard NASA's Compton Gamma-Ray Observatory between April 1991 and March 1993. Although astronomers agree that classes of gamma ray bursts exists, they do not agree on a specific class structure.

The Landsat Image Dataset.
The dataset contains pixels representing a digitized satellite image of a portion of the earth's surface. Each instance has been classified into one of fifteen categories. Because of the large number of individual classes, classification accuracy is affected by model-specific parameter settings.

The temperature Dataset.
The dataset offers the normal average January minimum temperature in degrees Fahrenheit for 56 U.S. cities. City latitude and longitude values are also provided. All attributes are numeric.
Miscellaneous

The Titanic Dataset.
The dataset contains 2201 instances where each instance describes attributes of an individual passenger or crew member aboard the Titanic. The output attribute indicates whether the passenger or crew member survived.