Learning from ambiguously labeled examples
Intelligent Data Analysis - Selected papers from IDA2005, Madrid, Spain
| Hi-index | 0.00 |
This paper introduces a framework of experienced-based decision making as an extension of case-based decision making, a recently proposed alternative to expected utility theory. In experienced-based decision making, an agent faced with a new decision problem acts on the basis of experience gathered from previous problems in the past, either through predicting the utility of potential actions or through establishing a direct relationship between decision problems and appropriate actions. In the paper, a realization of the latter approach in the form of “satisficing” decision trees is proposed.