人工智能培训

搜索

人工智能论文:特征选择的谱单纯理论及其在基因组学中的应用(Spectral Simplicial Theory for Feature Selection an

[复制链接]
xbgzs2010 发表于 5 天前 | 显示全部楼层 |阅读模式
xbgzs2010 5 天前 125 0 显示全部楼层
人工智能论文:特征选择的谱单纯理论及其在基因组学中的应用(Spectral Simplicial Theory for Feature Selection and Applications to  Genomics)现代数据集的规模和复杂性以及与测试大量假设相关的限制强调了对特征选择方法的需求。光谱技术根据其与基础度量结构的一致性程度对特征进行排名,但是它们基于当前图形的公式限制了它们对点特征的适用性。我们扩展了用于特征选择的谱方法来抽象单纯复形,并提出了一个可应用于2点和更高阶特征的通用框架。组合拉普拉斯分数考虑了数据所涵盖的拓扑,并在点特征的情况下降低到普通的拉普拉斯分数。我们通过几个应用于基因表达分析和多模式基因组数据的实例证明了光谱单纯形法用于特征选择的效用。我们的结果提供了拓扑数据分析和多种学习方法的统一视角。
The scale and complexity of modern data sets and the limitations associatedwith testing large numbers of hypotheses underline the need for featureselection methods.Spectral techniques rank features according to their degreeof consistency with an underlying metric structure, but their currentgraph-based formulation restricts their applicability to point features.Weextend spectral methods for feature selection to abstract simplicial complexesand present a general framework which can be applied to 2-point andhigher-order features.Combinatorial Laplacian scores take into account thetopology spanned by the data and reduce to the ordinary Laplacian score in thecase of point features.We demonstrate the utility of spectral simplicialmethods for feature selection with several examples of application to theanalysis of gene expression and multi-modal genomic data.Our results provide aunifying perspective on topological data analysis and manifold learningapproaches.人工智能论文:特征选择的谱单纯理论及其在基因组学中的应用(Spectral Simplicial Theory for Feature Selection and Applications to  Genomics) dDsqiM6ZPIVD66P9.jpg
URL地址:https://arxiv.org/abs/1811.03377     ----pdf下载地址:https://arxiv.org/pdf/1811.03377    ----人工智能论文:特征选择的谱单纯理论及其在基因组学中的应用(Spectral Simplicial Theory for Feature Selection and Applications to  Genomics)
回复

使用道具 举报

您需要登录后才可以回帖 登录 | 立即注册

本版积分规则 返回列表 发新帖

xbgzs2010当前离线
新手上路

查看:125 | 回复:0

快速回复 返回顶部 返回列表