摘要
针对逻辑回归分类模型,提出基于高维数据的集成逻辑回归分类算法,该算法随机抽取多个特征集,并针对各个特征集构建多个回归模型。并最终针对多个逻辑回归模型结果,利用集成学习方法进行最终预测。实验结果表明,集成逻辑回归分类算法具有很高的预测精度,与传统算法相比有明显的提高。
In this paper, focusing on logistic regression model, we propose ensemble logistic regression classification model based on high dimensional data. We selected features from the whole features set, and we predict the classification results with ensemble learning method. The experimental results show that the prediction accuracy of ensemble logistic regression classification algorithm is high, and has obvious promotion comparing with traditional algorithm.
出处
《科技通报》
北大核心
2013年第12期64-66,共3页
Bulletin of Science and Technology
基金
江苏省农委重大攻关项目(2130109)
关键词
高维数据
集成学习
逻辑回归
分类算法
随机子空间
high dimensional data
ensemble learning
logistic regression
classification algorithm,random subspace