摘要
由于乳腺X光图像的复杂性,直接从图像中看出肿瘤及其良、恶性质是很困难的,因此建立高效的肿瘤自动诊断系统非常必要。文中将关联规则分类器和粗糙集理论相结合构造了增强关联规则分类器(EAC),应用于乳腺X光图像分类。实验结果表明,EAC的分类精确度可达到77.48%,比单独使用关联规则的分类精确度(69.11%)要高近10%,同时规则数也明显减少。
Purpose. There exist only several methods based on data mining for classifying mammographic images . We present a new method, also based on data mining, that we believe is better than existing ones. In the full paper, we explain our new method in detail; in this abstract, we just list the two topics of our explanation ： （1） pretreating images and extracting their features ; （2） enhanced associative classifier （EAC）, whose subtopics are rough sets theory, associative rule, and algorithm for EAC. The experimental results, given in detail in Table 1 in the full paper, show that this EAC can get 77.48% classification accuracy which is higher than the 69. 11% obtained by Ref. 1 with associative classifier; furthermore the number of rules is much fewer than that of Ref. 1.
出处
《西北工业大学学报》
EI
CAS
CSCD
北大核心
2006年第3期401-404,共4页
Journal of Northwestern Polytechnical University
基金
国家自然科学基金(60373108
60573096)资助
关键词
增强关联规则分类器
粗糙集理论
乳腺X光图像
enhanced associative classifier （EAC）, rough sets theory, mammographic image
