摘要
在数据挖掘的许多实际应用中,在进行准确分类(classification)的同时,按照分类的可能性大小进行排序(ranking)日益显得重要。许多分类算法在设计时只考虑分类的准确性,未考虑对分类的可能性进行度量,因而无法用于排序(rank-ing)任务。本文提出了一种新的基于遗传算法的数据挖掘方法,在产生分类规则的同时,对分类的可能性进行度量。实验证明该算法是可行的。
In real-world data mining applications, an accurate ranking is same important to an accurate classification. A few algorithms are developed only for obtaining accurate classification, when it comes to determining the likelihood of each classification made for Ranking, many of them are not designed with such prupose in mind. The paper proposes a genetic algorithm for data mining, which can meet the demands mentioned above. The scheme is proved to be practicable.
出处
《计算机与现代化》
2006年第10期32-34,37,共4页
Computer and Modernization
关键词
数据挖掘
分类
排序
遗传算法
data mining
classification
ranking
genetic algorithm
