Lazy Averaged One-Dependence Estimators
Lazy Averaged One-Dependence Estimators
复制标题
DOI：
10.1007/11766247_44
复制
发表时间：
2006-06
期刊：
影响因子：
--
通讯作者：
Liangxiao Jiang;Harry Zhang
中科院分区：
文献类型：
--
作者：
Liangxiao Jiang;Harry Zhang
朴素贝叶斯是一种基于条件独立假设的概率分类模型。然而，在许多实际应用中，这一假设经常被违反。针对这一事实，研究人员已经做出了大量的努力，通过削弱条件独立假设来提高朴素贝叶斯的准确性。最近的工作是平均单依赖估计量（AODE）[15]，它表现出良好的分类性能。本文通过对AODE算法的扩展，提出了一种新的懒惰学习算法--Lazy Averaged One-Dependence Estimators，简称LAODE。对于给定的测试实例，LAODE首先根据每个训练实例与测试实例的相似性，对训练数据进行扩展，增加一些训练实例的副本（克隆），然后利用扩展后的训练数据构建AODE分类器对测试实例进行分类。我们在Weka系统[16]中实验测试了我们的算法，使用Weka [17]推荐的全部36个UCI数据集[11]，并将其与朴素贝叶斯[3]，AODE [15]和LBR [19]进行比较。实验结果表明，LAODE的性能显着优于所有其他用于比较的算法。
Naive Bayes is a probability-based classification model based on the conditional independence assumption. In many real-world applications, however, this assumption is often violated. Responding to this fact, researchers have made a substantial amount of effort to improve the accuracy of naive Bayes by weakening the conditional independence assumption. The most recent work is theAveraged One-Dependence Estimators(AODE) [15] that demonstrates good classification performance. In this paper, we propose a novel lazy learning algorithmLazy Averaged One-Dependence Estimators, simply LAODE, by extending AODE. For a given test instance, LAODE firstly expands the training data by adding some copies (clones) of each training instance according to its similarity to the test instance, and then uses the expanded training data to build an AODE classifier to classify the test instance. We experimentally test our algorithm in Weka system [16], using the whole 36 UCI data sets [11] recommended by Weka [17], and compare it to naive Bayes [3], AODE [15], and LBR [19]. The experimental results show that LAODE significantly outperforms all the other algorithms used to compare.