【作者】 令狐大智； 武新丽； 王东红；

【Author】 LINGHU Dazhi1,WU Xinli2,WANG Donghong3 (1.School of Business,Guangxi University,Nanning 530004,China; 2. School of Xingjian Art & Science,Guangxi University,Nanning 530005,China; 3. Department of Business & Administration,Guangxi College of Finance and Economics,Nanning 530003,China)

【机构】 广西大学商学院； 广西大学行健文理学院； 广西财经学院工商管理系；

【摘要】 针对基于约束的模糊数据归类算法的不足,构建模式自学习调节模型,并提出基于约束规则的模式微调算法CIP、基于自学习的分类规则优化更新算法OCRS。新算法基于模式间相关性、距离和支持度等因素,建立模式自主更新标准和算法协调机制。实验研究表明:新算法在准确度相同的情况下,增加了算法的识别率、自学习能力和鲁棒性。更多还原

【Abstract】 In order to overcome the shortcoings in restraint-based fuzzy data classification algorithm,a model of pattern self-learning is developed in this study. Also,a constraint-based inching pattern algorithm (CIP) and optimized classification rule algorithm based on self-learning (OCRS) are proposaed. The new algorithms incorporate the standard of pattern self-update and the coordination mechanism which are based on various factors,such as models correlation,distance and support etc. An experimental study shows that the new algorithms enhance the ability of self-learning and the robustness.更多还原