基于WEKA的序列最小化算法的改进研究 |
| |
引用本文: | 王朝辉,黎鑫. 基于WEKA的序列最小化算法的改进研究[J]. 工业控制计算机, 2012, 25(8): 81-82,84 |
| |
作者姓名: | 王朝辉 黎鑫 |
| |
作者单位: | 武汉科技大学计算机科学与技术学院,湖北武汉,430065 |
| |
摘 要: | 数据挖掘算法中的支持向量机算法,在通过若干学者的改进研究后,有一种改进算法即序列最小化算法主要应用于小样本数据集的分类,且分类效果较好,但在训练大规模数据集时,用时长、所需存储空间大,挖掘效率低。针对这一缺陷,通过改变存储策略改进该算法,在WEKA这个软件平台下,在保证分类正确率的前提下,缩短了训练时间,缩减了大量的存储空间,大大地提高了算法的效率,使其更加适应大规模数据集的训练。
|
关 键 词: | 数据挖掘 序列最小化算法 大规模数据集 挖掘效率 WEKA |
Research on Sequential Minimal Optimization Algorithm |
| |
Abstract: | Support Vector Machine algorithm is one of data mining algorithms.And after some scholars' research,there is a new algorithm that is Sequential Minimal Optimization Algorithm(SMO) which is usually used in the small sample datasets classification,and classification is better than other methods,but when training large-scale datasets,it takes long time to train,and needs high storage space.The mining efficiency is lower than small datasets.In response to this defect,improving the algorithm by changing the storage strategy in WEKA software platform. |
| |
Keywords: | data mining SMO large dataset digging efficiency WEKA |
本文献已被 CNKI 维普 万方数据 等数据库收录! |
|