示例学习扩张矩阵的粗糙计算

为了获得高效率和更简洁的知识，一些启发式算法被提出用于基于扩张矩阵理论的示例学习研究。该文基于粗集理论研究扩张矩阵的示例学习问题，并应用遗传算法获取示例学习中的最优概念。实验结果表明该方法是有效的。提出了粗集理论下的几个新概念，如：必要选择子，核选择子集，约简选择子集和所产生复合的评价指标。

Rough Computation of Extension Matrix for Learning from Examples

To achieve a higher efficient learning rate and to have a simplified representation, extension matrix theory has been applied in several heuristic algorithms for learning practically useful rules from examples. This paper presents a rough set theory based computational approach using the extension matrix theory for learning from examples. The implemented algorithm applied a GA method to perform optimal learning from given examples. Our experimental results reveal that this approach is potentially useful. Several new concepts including indispensable selector, core set of selectors, reduced set of selectors as well as accuracy, coverage and simplicity for evaluating a complex, are also proposed.