Research

Research Interests:

Artificial Intelligence
Case-based reasoning
Data mining
Machine Learning
Cognitive Science
Integrated approaches to reasoning and learning
Use of CBR for Law Enforcement

Research Project Experience:

Case-Editing for Numeric Prediction (detecting noisy or anomalous cases)(with T.J. Highley, La Salle University)
- Data used (dataset created for a project below) has been submitted to the machine learning archive at UCI (Cal - Irvine) UCI Machine Learning Repository. ("Communities and Crime" and "Communities and Crime Unnormalized")
Automatic Decomposition of Entity Relationship Diagrams (with M. Tavana and P. Joglekar, La Salle University)
Clustering for Cellular Manufacturing (with M. Tavana and P. Joglekar, La Salle University)
Predicting Community Crime Rates using Case-Based Reasoning (with C. Line, Rowan University). The dataset from this project has been donated to the UCI Machine Learning Repository. ("Communities and Crime")
Predicting Repeat Crime Victimizations using Case-Based Reasoning (with C. Line, Rowan University)
Simulating Resource Allocation for R & D Projects (with A. Kumaraswamy and A. Baveja, Rutgers University)
Providing Technical Assistance to Local Police Departments using Artificial Intelligence Case-Based Retrieval Techniques (software Crime X-windows Similarity (CXS) (with A. Baveja, Rutgers University) See sample screen dump.
See draft journal paper (postscript) (some gremlin mucked up the pagebreaks on the way to postscript from Word; I blame Bill Gates)
See short paper published in "The Forum" (news and research for southern New Jersey and Delaware Valley Public Administrators. (postscript)
or See an abstract of my 1997 INFORMS law enforcement CBR talk. ) Also, see beginnings of a web version (non-functional so far.)
Alter-Ego - Agent for filtering e-mail, using user model, world model, and content of messages. (with Beth Adelson) See CBR 98 Workshop (at AAAI) paper (postscript)
CECELIA - Chemistry Education Tutor -- Inducing Self-Explanations using a Case-Based Tutor (with Susan Phillips) Click here to see an abstract of my CECELIA talk.
See ICCBR 97 paper (postscript) (unfortunately some figures (screen dumps from CECELIA) were cut-and-pasted on paper and do not appear here)
CELIA -- Apprenticeship Learning -- learning to diagnose by observing and interacting with an expert.
CELIA was a large project with multiple major issues addressed:
- Diagnosis treated as iterative procedural task, rather than classification.
- Case Acquisition for Case-based Reasoning
- Case Representation for Diagnostic Case-based Reasoning
- Knowledge Acquisition Planning -- system decides what information to pursue and how to acquire it. Used especially for learning indices and filters for parts of cases.
- Qualitative component-centered domain model for automobile diagnosis
- Making use of both cases and model knowledge in a flexible manner for diagnosis
- Case-based empirical adjustment of feature importance. This improves similarity matching for case retrieval.
  See CBR 94 Workshop (at AAAI) paper (postscript)
  Click here to see the first (of 3) data files used with CELIA
  Click here to see the second (of 3) data files used with CELIA
  Click here to see the third (of 3) data files used with CELIA
"Greedy" assignment of students to groups based on time schedules, experience, and project preference

Click here to see a list of my refereed publications (+ my dissertation).

Resources:

UTCS Machine Learning Research Group some popular programs - useful for ML class demos
David Aha David W. Aha's Home Page Keeper of many AI resources including extensive people list. The below were moved off his Navy site, and can be found as links from this page.
- Home Pages of ML & CBR Folks
- David W. Aha: Machine Learning Page
- David W. Aha: CBR Resources

Data

Seemingly Broken Links:

Last updated August 15, 2017