...
首页> 外文期刊>International Journal of Computer Science and Technology >Efficient Technique for Classifying High-Dimensional Data
【24h】

Efficient Technique for Classifying High-Dimensional Data

机译:分类高维数据的有效技术

获取原文
           

摘要

Grouping issues in high dimensional data with few perceptions are ending up more typical particularly in microarray data.The two unique sorts of online feature selection tasks: 1) OFS by learning with full sources of information, and 2) OFS by learning with incomplete sources of information. Assume in first task that the learner can access all the features of training instances, and the goal is to efficiently identify a fixed number of relevant features for accurate prediction. In the second task, consider a more challenging scenario where the learner is allowed to access a fixed small number of features for each training instance to identify the subset of relevant features. This work proposes a new estimation measure Q-statistic that includes the solidity of the selected feature subset in addition to the estimate accuracy. Then propose the Booster of an FS algorithm that boosts the value of the Q-statistic of the algorithm applied. Empirical studies based on synthetic data and 14 microarray data sets show that Booster boosts not only the value of the Q-statistic but also the estimate accuracy of the algorithm applied unless the data set is intrinsically difficult to predict with the given algorithm.Full Paper
机译:对高维数据几乎没有感知的问题分组尤其在微阵列数据中变得更加典型。两种独特的在线特征选择任务:1)通过学习完整的信息资源进行OFS,以及2)通过学习不完整的数据来源进行OFS信息。假设第一个任务是学习者可以访问训练实例的所有功能,并且目标是有效地识别固定数量的相关功能以进行准确的预测。在第二个任务中,考虑一个更具挑战性的场景,其中允许学习者为每个训练实例访问固定的少量特征,以标识相关特征的子集。这项工作提出了一种新的估计量度Q统计量,该统计量除了估计精度外还包括所选特征子集的可靠性。然后提出一种FS算法的Booster,该Booster可以提高所应用算法的Q统计量的值。基于合成数据和14个微阵列数据集的经验研究表明,Booster不仅可以提高Q统计量的值,还可以提高所应用算法的估计精度,除非使用给定算法本质上难以预测该数据集。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号