用人工鱼群算法自动确定三支决策阈值

Applying Artificial Fish Swarm Algorithm to Automatically Determine Thresholds in Three- way Decision- theoretic Rough Set Model

传统的三支决策粗糙集模型需要设置合适的阈值,需要运用该模型的人员具备相关的专业知识和经验,这阻碍了该模型在实际中的应用。针对此不足,本文提出用人工鱼群算法来自动生成阈值,而不需要先验知识。以样本的条件概率作为解空间,以决策风险最小化为目标,利用人工鱼群算法,能有效地从数据中学习到三支决策粗糙集模型所需要的阈值,使得风险损失最小。在部分UCI数据集上的实验表明,该算法在运行时间上和利用学习到的阈值构建的分类器的分类性能都明显优于自适应算法。

The traditional three- wany decision rough set model needs to set up appropriate threshold. It requires the user of the model to have the relevant professional knowledge and experience,which hinders the application of the model in practice. To solve this problem,the artificial fish swarm algorithm is proposed to generate the threshold automatically,without requiring priori knowledge. Taking the conditional probability of sample as the target,using the artificial fish swarm algorithm,it can effectively learn from the data to the threshold required by the three- wany decision rough set model. It can make the risk loss minimum.The experimental result in part of UCI data sets shows that the algorithm run much faster than the adaptive learning parameters algorithm,and a three- decision- making classifier was built by using the threshold and this classifier can also make classifier better.