Mixed-initiative problem solving with decision trees

Authors:
David Mcsherry
Affiliations:
School of Computing and Information Engineering, University of Ulster, Coleraine, Northern Ireland, UK BT52 1SA
Venue:
Artificial Intelligence Review
Year:
2007

Citing 9
Cited 0

Conversational Case-Based Reasoning

Applied Intelligence
ExpertGuide: A Conversational Case-Based Reasoning Tool for Developing Mentors in Knowledge Spaces

Applied Intelligence
Interactive Case-Based Reasoning in Sequential Diagnosis

Applied Intelligence
Acquiring Customers' Requirementsin Electronic Commerce

Artificial Intelligence Review
Mixed-Initiative Interaction

IEEE Intelligent Systems
Induction of Decision Trees

Machine Learning
Metacognition in computation: a selected research review

Artificial Intelligence
Advances in conversational case-based reasoning

The Knowledge Engineering Review
A personalized system for conversational recommendations

Journal of Artificial Intelligence Research

Quantified Score

Hi-index	0.00

Visualization

Abstract

We present a mixed-initiative approach to problem solving based on decision trees in which the user can answer unknown to any question she is asked by the intelligent system, or answer questions anywhere in the decision tree without waiting to be asked. Also in contrast to the traditional decision-tree approach, more than one of the rules in a decision tree may contribute to the solution of a problem for which the user is unable to provide a complete description. As shown by our results, increased coverage of incomplete problem descriptions is an important benefit for some decision trees. However, a potential risk in allowing a problem-solving dialogue to continue when relevant test results are not available is that no solution may be possible no matter what additional information the user can provide. This problem is addressed in our approach by using meta-level reasoning to recognize when no solution is possible.