言葉ã¨æ¦‚念をç´ä»˜ã‘る〜Googleã«ã‚ˆã‚‹ã‚»ãƒžãƒ³ãƒ†ã‚£ãƒƒã‚¯Webå†æ§‹ç¯‰ - éšæ‰€ä½œä¸»éŒ²ã‚»ãƒžãƒ³ãƒ†ã‚£ãƒƒã‚¯Webã¨ã„ã†è¨€è‘‰ã®æ„味ã«è¸ã¿è¾¼ã‚“ã 文書データ処ç†ãŒè©±é¡Œã«ãªã£ã¦10å¹´ãらã„経ã¤ã¨æ€ã„ã¾ã™ãŒã€è¯ã€…ã—ãサービスã¨ã—ã¦ä½¿ã‚ã‚Œã¦ã„ã‚‹ã¨ã¯è€³ã«ã—ã¾ã›ã‚“ã—ã€ãƒ¦ãƒ¼ã‚¶ãƒ¼ã¨ã—ã¦ã‚‚ãã®æ©æµã«é ã‹ã£ã¦ã‚‹æ„Ÿè¦šã¯ã‚ã‚Šã¾ã›ã‚“。やã¯ã‚Šè¡¨è¨˜ã‚†ã‚Œã‚„åŒç¾©èªžã€æ„味階層ã«é–¢ã™ã‚‹è†¨å¤§ãªãƒ«ãƒ¼ãƒ«ä½“系（オントãƒã‚¸ãƒ¼ï¼‰ã‚’完璧ã«ä½œã‚Šä¸Šã’ãŸã‚Šã€ã‚³ãƒ³ãƒ†ãƒ³ãƒ„ã«å¯¾ã™ã‚‹ãƒ¡ã‚¿ãƒ‡ãƒ¼ã‚¿ä»˜ã‘ãŸã‚Šã¨ã„ã†äººæ‰‹ã®ãƒœãƒˆãƒ«ãƒãƒƒã‚¯ãŒè§£æ¶ˆã•ã‚Œãªã‹ã£ãŸã“ã¨ãŒåŽŸå› ãªã®ã§ã—ょã†ã‹ã€‚ãã‚“ãªåœæ»žã—ãŸã‚¨ãƒªã‚¢ã«ã€GoogleãŒã„ã¤ã‚‚通り少々暴力的（？）ãªæ–¹æ³•ã§ä¸€çŸ³ã‚’投ã˜ã‚‹æˆæžœã‚’公開ã—ãŸã‚ˆã†ã§ã™ã€‚ From Words to Concepts and Back: Dictionaries for Linking Text, Entities and Ideas [Google Research Blog] ä¸èº«ã¯ã„ãŸã£ã¦ã‚·ãƒ³ãƒ—ルã§ã€è‹±èªžç‰ˆWikipedia記事1ページをã
ç•°ãªã‚‹ã€Œã—ょã†ã¶ã€ を見分ã‘ã‚‹Google ã®åŒç¾©èªžã‚·ã‚¹ãƒ†ãƒ 
Google Labs - Books Ngram Viewer
Latent Dirichlet Allocation (LDA) and Google's Rankings are Remarkably Well Correlated
Lightly Supervised Learning of Text Normalization: Russian Number Names Abstract Most areas of natural language pro- cessing today make heavy use of au- tomatic inference from large corpora. One exception is text-normalization for such applications as te
SVMã«ã‚ˆã‚‹äºˆæ¸¬å¤‰æ› - nokunoã®æ—¥è¨˜
WWW2009
WWW2009
Average-Class-plot_diff-seeds.eps
ACL Anthology
Reading the Markets: Forecasting Public Opinion of Political Candidates by News Analysis Kevin Lerman Dept. of Computer Science Columbia University New York, NY USA 
Google Japan Blog: 大è¦æ¨¡æ—¥æœ¬èªž n-gram データã®å…¬é–‹
{{#tags}}- {{label}}
{{/tags}}