Generating estimates of classification confidence for a case-based spam filter
ICCBR'05 Proceedings of the 6th international conference on Case-Based Reasoning Research and Development
Navigating through case base competence
ICCBR'05 Proceedings of the 6th international conference on Case-Based Reasoning Research and Development
| Hi-index | 0.01 |
Case-Based Reasoning (CBR) is a learning approach that solves current situations by reusing previous solutions that are stored in a case base. In the CBR cycle the reuse step plays an important role into the problem solving process, since the solution for a new problem is based in the available solutions of the retrieved cases. In classification tasks a trivial reuse method is commonly used, which takes into account the most frequently solution proposed by the set of retrieved cases. We propose an alternative reuse process; we call confidence-reuse method, which make a qualitative assessment of the information retrieved. This approach is focused on measuring the solution accuracy, applying some confidence predictors based in a k-NN classifier with the aim of analyzing and evaluating the information offered by the retrieved cases.