Establishment of Statistical Estimation Principle for Super HighDimensional Data and Its Application to Large Scale Data Mining

超高维数据统计估计原理的建立及其在大规模数据挖掘中的应用

• 批准号: 22300054
• 负责人: WASHIO Takashi
• 金额: $ 11.07万
• 依托单位: Osaka University
• 依托单位国家: 日本
• 项目类别: Grant-in-Aid for Scientific Research (B)
• 财政年份: 2010
• 资助国家: 日本
• 起止时间: 2010 至 2012
• 项目状态: 已结题
• 来源: https://kaken.nii.ac.jp/grant/KAKENHI-PROJECT-22300054/
• 关键词: 知能発見; データマイニング; 統計的推定; 超高次元データ; ビッグデータ; 高次元データ; 次元の呪い; 粒子フィルタ; 統計数理; 確率密度関数; データ分布

Upon analysis of dimensionality curse, we characterized “hyper-sphere concentration effect”, “probability concentration effect” and “sparsity effect”of super high dimensional data, and proposed an accurate and robust estimation method against the former two effects.

通过对维数诅咒的分析，我们刻画了超高维数据的“超球集中效应”、“概率集中效应”和“稀疏效应”，并针对前两种效应提出了一种准确鲁棒的估计方法。

项目成果

Use of Prior Knowledge in a Non-Gaussian Method for Learning Linear Structural Equation Models

• DOI: 10.1007/978-3-642-15995-4_28
• 发表时间: 2010-09
• 作者: Takanori Inazumi;Shohei Shimizu;T. Washio

Prismatic Algorithm for Discrete D.C.Programming Problem

离散直流规划问题的棱柱算法

• 发表时间: 2011
• 作者: Yoshinobu Kawahara;Takashi Washio
• 通讯作者: Takashi Washio

Separation of stationary and non-stationary sources with a generalized eigenvalue problem

• DOI: 10.1016/j.neunet.2012.04.001
• 发表时间: 2012-09-01
• 期刊: NEURAL NETWORKS
• 影响因子: 7.8
• 作者: Hara, Satoshi;Kawahara, Yoshinobu;Yumoto, Kiyohumi
• 通讯作者: Yumoto, Kiyohumi

Learning a common substructure of multiple graphical Gaussian models

• DOI: 10.1016/j.neunet.2012.11.004
• 发表时间: 2013-02-01
• 期刊: NEURAL NETWORKS
• 影响因子: 7.8
• 作者: Hara, Satoshi;Washio, Takashi
• 通讯作者: Washio, Takashi

Analyzing relationships among ARMA processes based on non-Gaussianity of external influences

• DOI: 10.1016/j.neucom.2011.02.008
• 发表时间: 2011-06
• 期刊: Neurocomputing
• 影响因子: 6
• 作者: Yoshinobu Kawahara;Shohei Shimizu;T. Washio