Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Abstract Information for Xavier Celebration of Student Research and Creative Activity April 4, 2005 Abstract Instructions Please indicate type of presentation: ___ poster _X_ 15 min. oral presentation Is special audio visual equipment necessary (ex. PowerPoint)? ___no _X_yes If yes, please call Mary Alice Vetter at 745-3101 to make arrangements. Font: The required font is New Times Roman, Font size 10. The abstract title should be typed in capital letters. Only the title line should be in capital letters. Immediately following the title line should be a line that contains the Authors' Names (first name, middle initial, last name.) The faculty mentor should be the last author listed and this name should be placed in parenthesis. The next line should be the department name. The abstract should follow starting on the next line and there should be no indentation. All information must fit into the space provided. No changes are to be made to the font size or margins of this page to accommodate your abstract. E-mail an attached file copy of this form with the appropriate changes to the content of the example abstract below to Dr. Barbara Hopkins at [email protected]. Also print a final copy of the form and mail this paper copy to Mary Alice Vetter, ML 4543. Final day for submission of abstracts is March 7. See second attachment for examples. Use only the area below to type your abstract. Change content but do not change font or margins. MINING DISJUNCTIVE ASSOCIATION RULES USING GENETIC PROGRAMMING Michelle A. Lyman (Dr. Gary Lewandowski) Department of Mathematics and Computer Science Machine learning techniques can be used to mine trends in data sets. This paper reports findings mined from data collected in a card sorting exercise performed by beginning Computer Science students and their professors. Participants sorted 26 cards with Computer Science-related concepts written on them into groups based on their understanding of the concepts. An application was developed to mine overarching trends in the way the participants sorted the cards. These trends were expressed in the form of disjunctive association rules. The application randomly generated association rules and then used genetic programming techniques to evolve them into rules with high support and confidence. When the card sorting data was segmented according to the achievement level of the participant, differences in the number of association rules produced by each group could be observed. This information, along with the actual association rules, can be used to present Computer Science concepts more effectively to beginning Computer Science students.