粗糙集属性约简算法在图像分类研究的应用

基于贝叶斯粗糙集,引入贝叶斯区分矩阵,采用属性的出现频率与属性的长度作为启发因素,并以此给出了贝叶斯粗糙集属性约简的另外一种算法,最后提出了一种基于颜色特征的图像分类模型及其分类算法。用该方法进行图像资源的分类,克服了经典粗糙集不宜处理带有噪声的数据和决策表不协调的分类问题的缺陷,同时又大大简化分类规则,且形成的规则集便于用户理解。完善了近似空间的概念。实验结果表明在处理决策表不协调的图像分类问题,贝叶斯粗糙集方法性能良好,分类准确和高效。

The Bayesian Rough Set Based Approach for Image Classification

Based on Bayesian Rough Set, introducing the distinction between Bayesian matrix, frequency of use of properties and attributes as a stimulating factor in the length, the article gives the Bayesian rough set attribute reduction algorithm for another, and finally put forward a Based on the characteristics of a color image classification model and its classification algorithm. With this method, the classification of image resources to overcome the classical rough set not to deal with noise data and decision-m...