一种基于粗糙集的味觉信号识别方法

A recognition method of taste signals based on rough set

提出一种基于粗糙集的味觉信号识别方法.该算法运用粗糙集技术,在决策规则生成过程中,充分考虑数据集中各属性的重要度,并动态对其进行更新.由于决策过程中不断更新属性重要度,保证了每次将重要度最高的属性加入决策规则集,进而保证了决策系统的约简.基于机器学习数据集UCI中的2个味觉信号数据winequality_white和winequality_red,算法采用十折交叉验证技术,独立进行10次实验,并与2个经典算法进行了对比.结果表明,本文算法的味觉信号识别正确率更高、更有效.

This is a recognition method of taste signals based on rough set. For the realization of computer taste, the identification of taste signals is also very important except developing taste sensor with high sensitivity. The algorithm of this paper bases on rough set technology. It fully considers attribute significances and updates them dynamically during the process of decision making. Because of constantly updating attribute significances,it ensures to add the attribute with highest significance to the rule set. So, the decision system is simple. Based on two machine learning data set UCI, winequality-white and winequality-red, the proposed algorithm adopts ten-fold cross validation technology to run ten times independently. And comparing with the two other classical algorithms, results show that the proposed algorithm is better and more effective than them on taste signal recognition rate.