摘要
在全面综述通用机器学习归纳分类算法的基础上,采用算法分类机制分析的方法,从预测精度、学习效率、健壮性等方面对决策树和规则归纳分类算法进行深入的分析和比较研究,为在不同的应用领域选择最优分类算法奠定了基础。由于RIPPER分类算法采用了重复增量裁减机制,所以在计算复杂性、分类精度、噪音数据适应性等方面都优于其它分类算法,更适用于入侵检测建模使用。
It summarizes the main features of decision tree learning algorithm and rule learning algorithm by in-depth analysis and comparison from all aspects such as prediction accuracy, learning efficiency and robustness. It is shown that RIPPER is superior to other algorithms in terms of complexity in computation, classified precision and noisy data adaptability because of its adoption of the repeated incremental reduction mechanism, and it is more suitable to the intrusion detection.
出处
《电子产品可靠性与环境试验》
2004年第6期72-75,共4页
Electronic Product Reliability and Environmental Testing
基金
山西省自然科学基金项目(20041047)资助。
关键词
机器学习
分类算法
入侵检测
machine learning
classification algorithm
intrusion detection
