BFC: Bat Algorithm Based Fuzzy Classifier for Medical Data Classification
BFC: Bat Algorithm Based Fuzzy Classifier for Medical Data Classification
复制标题
DOI：
10.1166/jmihi.2015.1428
复制
发表时间：
2015-06-01
影响因子：
--
通讯作者：
Selvi, M.
中科院分区：
文献类型：
--
作者：
Binu, D.;Selvi, M.
优化算法应用于模糊系统的各种目的，如隶属函数优化，系数优化，规则生成，规则选择等，在这里，我们合并蝙蝠算法与模糊分类器，有效地生成优化的规则和隶属函数。我们的分类器的关键贡献是（i）使用蝙蝠算法生成和选择优化的规则，（ii）简化隶属函数的设计和离散化过程，（iii）根据规则在学习数据中的出现频率制定适应度函数。提出的蝙蝠算法的模糊分类器进行定量和定性分析的性能比较。实验结果表明，当使用肺癌数据时，BFC达到了75.21%的准确率。此外，BFC对印度肝脏数据的准确率达到76.67%。
Optimization algorithms are applied on Fuzzy system for various purposes like membership function optimization, co-efficient optimization, rule generation, rule selection, etc. Here we amalgamate Bat algorithm with fuzzy classifier to generate optimized rules and membership functions effectively. The key contributions in our classifier are (i) generating and selecting optimized rules using bat algorithm, (ii) simplifying the design and discretizing process in membership function, (iii) formulating a fitness function based on frequency of occurrence of the rules in the learning data. The proposed Bat algorithm based fuzzy classifier is subjected to quantitative and qualitative analysis for performance comparisons. Experimental results demonstrate that BFC has achieved 75.21% accuracy when Lung cancer data is used. Moreover, BFC has accomplished 76.67% accuracy for Indian Liver data.