Large-scale tensor decomposition for high-level information processing

Research Project

Project/Area Number	25880028
Research Category	Grant-in-Aid for Research Activity Start-up
Allocation Type	Single-year Grants
Research Field	Intelligent informatics
Research Institution	National Institute of Informatics
Principal Investigator	HAYASHI Kohei 国立情報学研究所, ビッグデータ数理国際研究センター, 特任助教 (30705059)
Project Period (FY)	2013-08-30 – 2015-03-31
Project Status	Completed (Fiscal Year 2014)
Budget Amount *help	¥2,730,000 (Direct Cost: ¥2,100,000、Indirect Cost: ¥630,000) Fiscal Year 2014: ¥1,300,000 (Direct Cost: ¥1,000,000、Indirect Cost: ¥300,000) Fiscal Year 2013: ¥1,430,000 (Direct Cost: ¥1,100,000、Indirect Cost: ¥330,000)
Keywords	非負行列分解 / Twitter / トピック抽出 / ストリーミング処理 / テンソル分解 / 確率勾配法 / 関係データ解析 / ソーシャルネットワーク解析 / トピックモデル
Outline of Final Research Achievements	We proposed a streaming algorithm of non-negative matrix factorization for a sequence of matrices such as time series. We evaluated its performance by applying to the Twitter stream, which is one of the popular social network services. Our algorithm allows to handle all Japanese tweets in real-time, which tells us what topics are "hot" in that moment for every second.

Report

(3 results)