| Popis: |
Adding fuzziness to normal classification rules enables the rules to adapt to the real-life decision-making process. Besides, it also adds to the classification accuracy of the obtained model and the rules look more accurate and reasonable. Further improvement in classification accuracy can be achieved by discovering exceptions corresponding to these fuzzy rules. Fuzzy rules augmented with exceptions (censors) are termed as Fuzzy Censored Classification Rules (FCCRs) and such kind of rules are best at handling uncertainties like vagueness and ambiguity. These rules, being very efficient, have been widely used under exceptional circumstances. In this paper, we have investigated all the algorithms used in past for discovering FCCRs. Based on review of literature, we have also proposed possible modifications to existing algorithms and techniques. |
