|
| 面向不平衡数据的逻辑回归偏标记学习算法研究 |
| Partial label learning algorithm for data imbalance problem based on logistic regression |
| 投稿时间：2016-09-05 修订日期：2016-09-12 |
| DOI： |
| 中文关键词: 偏标记学习 数据不平衡 逻辑回归 阻尼牛顿法 |
| 英文关键词: partial label learning data imbalance logistic regression damped Newton method |
| 基金项目:：国家自然科学基金项目(61503058,61374170,61502074，U1560102)；辽宁省自然科学基金项目(2015020084)；中央高 校基本科研业务费专项资金资助项目（DC201501055，DC201501060201）. |
|
| 摘要点击次数: 481 |
| 全文下载次数: 0 |
| 中文摘要: |
| ：偏标记学习是近几年提出的新机器学习框架，已有的逻辑回归偏标记算法还没有解决
不平衡问题。本文建立了一种可以解决数据不平衡的逻辑回归模偏标记学习算法。基本思
想是在多元逻辑回归模型中定义新的似然函数以达到处理不平衡数据的目的。算法先根据
训练集中各个类别样本所占比例定义了一个新的似然函数，之后通过逼近和求导等数学手
段推导得到了能够求解的光滑的逻辑回归偏标记学习模型。在 UCI 数据集和真实数据集上
的仿真实验表明，所提算法在数据存在不平衡问题时提高了样本的平均分类精度。 |
| 英文摘要: |
| Partial label learning which is a new machine learning framework was proposed in recent years, but existing partial label learning algorithms based on logistic regression have not solved the problem of imbalance. A partial label learning algorithm of data imbalance is present based on logistic regression model. The basic idea is defining a new likelihood function in the multiple logistic regression models for dealing with imbalanced data. First, a new likelihood function is defined according to the proportion of each class sample in the training set, then the smooth partial label learning model is derived through derivation and approximation method. Simulation experiments on UCI data sets and real world data sets show that the proposed algorithm improves the average classification accuracy of sample. |
|
View Fulltext
查看/发表评论 下载PDF阅读器 |
| 关闭 |