[B! æ©Ÿæ¢°å¦ç¿’] nextbigthingã®ãƒ–ãƒƒã‚¯ãƒžãƒ¼ã‚¯

nextbigthing id:nextbigthing

æ©Ÿæ¢°å¦ç¿’ã«é–¢ã™ã‚‹nextbigthingã®ãƒ–ãƒƒã‚¯ãƒžãƒ¼ã‚¯ (31)

${{author_name}}$

{{{comment_expanded}}}

{{label}}

{{#is_bookmark}}ãƒªã‚¹ãƒˆ{{/is_bookmark}}{{^is_bookmark}}ãƒªãƒ³ã‚¯{{/is_bookmark}}

${{author_name}}$
{{author_name}}{{created}}
{{ #comment }}{{ comment }}{{ /comment }}
- {{ label }}

${{author_name}}$

{{{comment_expanded}}}

{{label}}

{{#is_bookmark}}ãƒªã‚¹ãƒˆ{{/is_bookmark}}{{^is_bookmark}}ãƒªãƒ³ã‚¯{{/is_bookmark}}

ã‚½ãƒ•ãƒˆãƒžãƒ¼ã‚¸ãƒ³SVM - äººå·¥çŸ¥èƒ½ã«é–¢ã™ã‚‹æ–å‰µéŒ²
å‰å›žï¼ˆ2010/5/2ï¼‰ã®ãƒãƒ¼ãƒ‰ãƒžãƒ¼ã‚¸ãƒ³SVMã§ã¯ã€ãƒ‡ãƒ¼ã‚¿ã«é‡ãªã‚ŠãŒã‚ã‚‹å ´åˆã€ä¸‹ã®ã‚ˆã†ã«ã¡ã‚ƒã‚“ã¨åˆ†é¡žå¢ƒç•Œã‚’æ±‚ã‚ã‚‰ã‚Œã¾ã›ã‚“ã§ã—ãŸã€‚ä»Šå›žã¯ã€é‡ãªã‚Šã®ã‚ã‚‹ã‚¯ãƒ©ã‚¹åˆ†å¸ƒã«å¯¾å¿œã§ãã‚‹ã‚ˆã†ã«æ‹¡å¼µã—ã¦ã¿ã¾ã™ã€‚ã“ã®ã‚ˆã†ãªSVMã¯ãƒãƒ¼ãƒ‰ãƒžãƒ¼ã‚¸ãƒ³SVMã«å¯¾ã—ã¦ã‚½ãƒ•ãƒˆãƒžãƒ¼ã‚¸ãƒ³SVMã¨å‘¼ã°ã‚Œã¾ã™ã€‚åˆ¥åã¨ã—ã¦C-SVMã¨ã‚‚å‘¼ã°ã‚Œã‚‹ã‚ˆã†ã§ã™ã€‚ PRMLã®7.1.1ã«ã‚ã‚‹ã‚ˆã†ã«ã€ãƒ‡ãƒ¼ã‚¿ã®èª¤åˆ†é¡žã‚’è¨±ã™ã‚ˆã†ã«SVMã‚’ä¿®æ£ã—ã¾ã™ã€‚ãƒãƒ¼ãƒ‰ãƒžãƒ¼ã‚¸ãƒ³SVMã§ã¯ã€ãƒ‡ãƒ¼ã‚¿ç‚¹ãŒãƒžãƒ¼ã‚¸ãƒ³å†…ï¼ˆ-1 < y < 1ï¼‰ã«çµ¶å¯¾ã«å…¥ã‚‰ãªã„ã“ã¨ã‚’å‰æã«ã—ã¦ã„ã¾ã—ãŸãŒã€ã‚½ãƒ•ãƒˆãƒžãƒ¼ã‚¸ãƒ³SVMã§ã¯ã€Œå…¥ã£ã¦ã—ã¾ã£ãŸã‚‚ã®ã¯ä»•æ–¹ãªã„ã€ã ãŒãƒšãƒŠãƒ«ãƒ†ã‚£ã‚’ä¸Žãˆã‚‹ï¼ã€ã¨å°‘ã—æ¡ä»¶ã‚’ç·©ã‚ã¾ã™ã€‚ ã¾ãšã€ã‚¹ãƒ©ãƒƒã‚¯å¤‰æ•°Î¶ï¼ˆã‚¼ãƒ¼ã‚¿ï¼‰ã‚’ãƒ‡ãƒ¼ã‚¿ã”ã¨ã«å°Žå…¥ã—ã¾ã™ã€‚ã‚¹ãƒ©ãƒƒã‚¯å¤‰æ•°ã¯ã€ãƒ‡ãƒ¼ã‚¿ãŒæ£ã—ãåˆ†é¡žã•ã‚Œã‹ã¤ãƒžãƒ¼ã‚¸ãƒ³å¢ƒç•Œä¸Šã¾ãŸã¯å¤–å´ã«ã‚ã‚‹å ´åˆã¯0ã€æ£ã—ãåˆ†é¡žã•ã‚Œã¦ã„ã‚‹ãŒãƒžãƒ¼ã‚¸ãƒ³å†…ã«ä¾µå…¥ã—ã¦ã—ã¾
nextbigthing 2013/09/03
python

æ©Ÿæ¢°å¦ç¿’
ãƒªãƒ³ã‚¯
SVM ã®ãƒãƒ¥ãƒ¼ãƒ‹ãƒ³ã‚°ã®ã—ã‹ãŸï¼ˆï¼’ï¼‰ - ã»ããç¬‘ã‚€
ã•ã¦ã€å‰å›žã¯äº¤å·®æ¤œè¨¼ã®èª¬æ˜Žã§çµ‚ã‚ã£ã¦ã—ã¾ã„ã¾ã—ãŸãŒã€ä»Šå›žã¯ã¡ã‚ƒã‚“ã¨ SVM ã®ãƒãƒ¥ãƒ¼ãƒ‹ãƒ³ã‚°ã®è©±ã‚’ã—ã¾ã™ã€‚ ãƒãƒ¥ãƒ¼ãƒ‹ãƒ³ã‚°ã®æ‰‹é †ã¨ã—ã¦ã¯ã€ ã‚°ãƒªãƒƒãƒ‰ã‚µãƒ¼ãƒã§å¤§é›‘æŠŠã«æ¤œç´¢ã™ã‚‹ã€‚ æœ€é©ãªãƒ‘ãƒ©ãƒ¡ãƒ¼ã‚¿ãŒã‚ã‚Šãã†ãªã¨ã“ã‚ã‚’çµžã£ã¦å†ã³ã‚°ãƒªãƒƒãƒ‰ã‚µãƒ¼ãƒã‚’è¡Œã†ã€‚ ã¨ã„ã†2æ®µéšŽã®ã‚°ãƒªãƒƒãƒ‰ã‚µãƒ¼ãƒã‚’è¡Œã„ã¾ã™ã€‚ 1æ®µéšŽç›®ï¼šã‚°ãƒªãƒƒãƒ‰ã‚µãƒ¼ãƒã§å¤§é›‘æŠŠã«æ¤œç´¢ã™ã‚‹ SVM ã®ãƒãƒ¥ãƒ¼ãƒ‹ãƒ³ã‚°ã¯ tune.svm() ã¨ã„ã†é–¢æ•°ã‚’ç”¨ã„ã¦è¡Œã„ã¾ã™ã€‚ ãƒãƒ¥ãƒ¼ãƒ‹ãƒ³ã‚°ã®ã‚„ã‚Šæ–¹ã¯ã€å˜ç´”ã«ã‚°ãƒªãƒƒãƒ‰ã‚µãƒ¼ãƒã‚’è¡Œã£ã¦ã„ã‚‹ã ã‘ã§ã™ã€‚ ãƒ‘ãƒ©ãƒ¡ãƒ¼ã‚¿ã®å€¤ã‚’ã„ã‚ã„ã‚å¤‰ãˆã¦ã¿ã¦ã€æ£ç”çŽ‡ã®ä¸€ç•ªã„ã„å€¤ã‚’ãƒ™ã‚¹ãƒˆãƒ‘ãƒ©ãƒ¡ãƒ¼ã‚¿ã¨ã—ã¦å‡ºåŠ›ã—ã¾ã™ã€‚ ãƒ—ãƒã‚°ãƒ©ãƒ ã¯ä¸‹è¨˜ã®ã‚ˆã†ã«ãªã‚Šã¾ã™ã€‚ gammaRange = 10^(-5:5) costRange = 10^(-2:2) t <- tune.svm(Species ~ ., data = iris, gamma=gammaRan
nextbigthing 2013/09/03
python

æ©Ÿæ¢°å¦ç¿’
ãƒªãƒ³ã‚¯
SVM ã®ãƒãƒ¥ãƒ¼ãƒ‹ãƒ³ã‚°ã®ã—ã‹ãŸï¼ˆï¼‘ï¼‰ - ã»ããç¬‘ã‚€
SVM ã®ãƒãƒ¥ãƒ¼ãƒ‹ãƒ³ã‚° SVM(Support Vector Machine) ã¯ã¿ãªã•ã‚“å¾¡å˜ã˜æ©Ÿæ¢°å¦ç¿’ã®æ‰‹æ³•ã§ã™ã€‚ SVM ã¯ãƒ‡ãƒ•ã‚©ãƒ«ãƒˆè¨å®šã§ãƒ¢ãƒ‡ãƒ«ã‚’ä½œã£ã¦ã‚‚ã—ã‚‡ã†ãŒãªã„ã§ã™ã€‚gamma ã¨ cost ã¨ã„ã†ãƒ‘ãƒ©ãƒ¡ãƒ¼ã‚¿ãŒã‚ã‚‹ã®ã§ã€ã“ã‚Œã‚‰ã®å€¤ã«æœ€é©å€¤ã‚’è¨å®šã—ãªã‘ã‚Œã°ãªã‚Šã¾ã›ã‚“ã€‚R ã® SVM ã® Help ã«ã‚‚ã“ã†æ›¸ã„ã¦ã‚ã‚Šã¾ã™ã€‚ Parameters of SVM-models usually must be tuned to yield sensible results! (è¨³) SVM ã§ã„ã„çµæžœå‡ºã—ãŸã‹ã£ãŸã‚‰ãƒãƒ¥ãƒ¼ãƒ‹ãƒ³ã‚°ã—ã‚ã‚ˆãªï¼ ã¨ã„ã†ã‚ã‘ã§ã€SVM ã®ãƒãƒ¥ãƒ¼ãƒ‹ãƒ³ã‚°ã®ã—ã‹ãŸã«ã¤ã„ã¦èª¬æ˜Žã—ãŸã„ã¨æ€ã„ã¾ã™ã€‚ äº¤å·®æ¤œè¨¼ ãŠã£ã¨ã€ãã®å‰ã«ã€äº¤å·®æ¤œè¨¼ã®è©±ã‚’ã—ãªã‘ã‚Œã°ãªã‚Šã¾ã›ã‚“ã€‚ SVM ãƒ¢ãƒ‡ãƒ«ã‚’ãƒãƒ¥ãƒ¼ãƒ‹ãƒ³ã‚°ã™ã‚‹éš›ã€äºŒã¤ã®ãƒ‘ãƒ©ãƒ¡ãƒ¼ã‚¿ã§ã‚°ãƒªãƒƒãƒ‰ã‚µãƒ¼ãƒã‚’ã—ã¾ã™ã€‚ ã™ãªã‚ã¡ã€ãƒ‘ãƒ©ãƒ¡ãƒ¼ã‚¿ã‚’ã„ã‚ã„ã‚å¤‰
nextbigthing 2013/09/03
python

æ©Ÿæ¢°å¦ç¿’
ãƒªãƒ³ã‚¯
2013.07.15 ã¯ã—ã‚™ãƒã‚šã‚¿lt scikit-learnã¦ã‚™å§‹ã‚ã‚‹æ©Ÿæ¢°å¦ç¿’
4. æ©Ÿæ¢°å¦ç¿’ã®ãƒ‘ãƒƒã‚±ãƒ¼ã‚¸ Rå‹¢ ï¬ e1071ï¼ˆSVMã€ã‚¯ãƒ©ã‚¹ã‚¿ãƒªãƒ³ã‚°ã¨ã‹ï¼‰ ï¬ gbmï¼ˆAdaBoostï¼‹å›žå¸°ãªã©ï¼‰ ï¬ kernlabï¼ˆSVMã€ã‚«ãƒ¼ãƒãƒ«ä¸»æˆåˆ†ã¨ã‹ï¼‰ ï¬ mboostï¼ˆboostingï¼‰ ï¬ nnetï¼ˆãƒ‹ãƒ¥ãƒ¼ãƒ©ãƒ«ãƒãƒƒãƒˆãƒ¯ãƒ¼ã‚¯ï¼‰ ï¬ randomForest ï¬ rpart etc. å®šç•ª ï¬ LIBSVM ãã®ä»– ï¬ weka (Java) ï¬ nltk (Python) æ–°èˆˆå‹¢åŠ› ï¬ Apache Mahout ï¬ Jubatus ï¬ scikit-learn ã¨ã„ã†å‹æ‰‹ãªå°è±¡ 5. æ©Ÿæ¢°å¦ç¿’ã®ãƒ‘ãƒƒã‚±ãƒ¼ã‚¸ Rå‹¢ ï¬ e1071ï¼ˆSVMã€ã‚¯ãƒ©ã‚¹ã‚¿ãƒªãƒ³ã‚°ã¨ã‹ï¼‰ ï¬ gbmï¼ˆAdaBoostï¼‹å›žå¸°ãªã©ï¼‰ ï¬ kernlabï¼ˆSVMã€ã‚«ãƒ¼ãƒãƒ«ä¸»æˆåˆ†ã¨ã‹ï¼‰ ï¬ mboostï¼ˆboostingï¼‰ ï¬ nnetï¼ˆãƒ‹ãƒ¥ãƒ¼ãƒ©ãƒ«ãƒãƒƒãƒˆãƒ¯ãƒ¼ã‚¯ï¼‰ ï¬ randomForest ï¬ rpar
nextbigthing 2013/08/31
æ©Ÿæ¢°å¦ç¿’

python
ãƒªãƒ³ã‚¯
Deep Learningã®æŠ€è¡“ã¨æœªæ¥
PFIã‚»ãƒŸãƒŠãƒ¼2013å¹´6æœˆ6æ—¥åˆ†ã§ã™ã€‚Deep Learningã®æŠ€è¡“çš„åŸºç¤Žã‹ã‚‰Bengioã®æ€ªã—ã’ãªå¦„æƒ³ã¾ã§ã€‚Read less
nextbigthing 2013/06/07
deep learning

æ©Ÿæ¢°å¦ç¿’
ãƒªãƒ³ã‚¯
ã‚³ãƒ³ãƒ”ãƒ¥ãƒ¼ã‚¿ã¨äººã€‚é è„³ã®å¯¾æ±ºã€‚åå‹è² ãŒç¶šãå°†æ£‹ãƒ»é›»çŽ‹æˆ¦ï¼ˆå¹³æž—ä¹…å’Œï¼‰ - ã‚¨ã‚ã‚¹ãƒ‘ãƒ¼ãƒˆ - Yahoo!ãƒ‹ãƒ¥ãƒ¼ã‚¹
æ¶™ã®å¼•ãåˆ†ã‘ãƒ—ãƒæ£‹å£«ã¨ã‚³ãƒ³ãƒ”ãƒ¥ãƒ¼ã‚¿ã®å¯¾æ±ºã€‚å°†æ£‹ãƒ»é›»çŽ‹æˆ¦ã€‚ ï¼”æœˆ13æ—¥ã€ç¬¬4å±€ãŒæ±äº¬ãƒ»åƒé§„ã‚±è°·ã®å°†æ£‹ä¼šé¤¨ã§è¡Œã‚ã‚ŒãŸã€‚ ä»Šã‚·ãƒªãƒ¼ã‚ºã®å¯¾æˆ¦æˆç¸¾ã¯ãƒ—ãƒæ£‹å£«ãŒï¼‘å‹ã€ã‚³ãƒ³ãƒ”ãƒ¥ãƒ¼ã‚¿ãŒï¼’å‹ã€‚ å¡šç”°æ³°æ˜Žä¹æ®µãŒè² ã‘ã‚Œã°ã€å›£ä½“æˆ¦äº”ç•ªå‹è² ãƒˆãƒ¼ã‚¿ãƒ«ã§ã‚‚ãƒ—ãƒæ£‹å£«å´ã®æ•—åŒ—ãŒæ±ºã¾ã£ã¦ã—ã¾ã†ã€‚æ´å²çš„å¤§ä¸€ç•ªã ã£ãŸã€‚å¡šç”°ä¹æ®µã®å¯¾æˆ¦ç›¸æ‰‹ã¯ã€æ˜¨å¹´ã€æ•…ãƒ»ç±³é•·é‚¦é›„å°†æ£‹é€£ç›Ÿä¼šé•·ã«å‹ã£ãŸã€Œãƒœãƒ³ã‚¯ãƒ©ãƒ¼ã‚ºã€ã®å¼·åŒ–ç‰ˆã€ã€ŒPuella Î±ï¼ˆãƒ—ã‚¨ãƒ©ã‚¢ãƒ«ãƒ•ã‚¡ï¼‰ã€ã€‚å¯Œå£«é€šç ”ç©¶æ‰€ã«æ‰€å±žã™ã‚‹ä¼Šè—¤è‹±ç´€æ°ãŒé–‹ç™ºã—ãŸã€‚ç¬¬21å›žä¸–ç•Œã‚³ãƒ³ãƒ”ãƒ¥ãƒ¼ã‚¿å°†æ£‹é¸æ‰‹æ¨©ã§å„ªå‹ã—ãŸã‚½ãƒ•ãƒˆã ã€‚ å¯¾å±€é–‹å§‹ã‹ã‚‰å¡šç”°ä¹æ®µã«å‹æ©Ÿã¯è¨ªã‚Œãªã‹ã£ãŸã€‚æ‚ªæ‰‹ã¯ä¸€æ‰‹ã‚‚æŒ‡ã—ã¦ã„ãªã„ã€‚ã«ã‚‚ã‹ã‹ã‚ã‚‰ãšã€ã‚³ãƒ³ãƒ”ãƒ¥ãƒ¼ã‚¿æœ‰åˆ©ã®å±€é¢ãŒç¶šã„ãŸã€‚çµ‚ç›¤ã«ã¯å¡šç”°ä¹æ®µãŒå¿…æ•—ã®å¤§å‹¢ã«ãªã£ãŸãŒã€ç²˜ã‚Šã«ç²˜ã£ã¦å¼•ãåˆ†ã‘ã«æŒã¡è¾¼ã‚“ã ã€‚çŽ‹ãŒæ•µé™£ã¾ã§é€²ã¿ã‚ã£ã¦ã€ãŠäº’ã„ã«å‹ã¡ãŒãªã„çŠ¶æ…‹ã€æŒå°†æ£‹ã¨ãªã£ãŸã€‚ 230æ‰‹ã®æ¿€ã—ã„æˆ¦ã„ãŒçµ‚ã‚ã‚Šã€è¦³æˆ¦è¨˜è€…
nextbigthing 2013/04/15
æ©Ÿæ¢°å¦ç¿’
ãƒªãƒ³ã‚¯
æ•°å¼ã‚’numpyã«è½ã¨ã—ã“ã‚€ã‚³ãƒ„
Tokyo.SciPy #2 ã«ã¦ç™ºè¡¨ã—ãŸã€æ•°å¼(ã‚ã‚‹ã„ã¯æ•°å¼å…¥ã‚Šã®ã‚¢ãƒ«ã‚´ãƒªã‚ºãƒ )ã‹ã‚‰å®Ÿè£…ã«è½ã¨ã™å ´åˆã€ä½•ã«æ°—ã‚’ã¤ã‘ã‚‹ã®ã‹ã€ã©ã†è€ƒãˆã‚Œã°ã„ã„ã®ã‹ã€ã¨ã„ã†ãŠè©±ã€‚ å¯¾è±¡ã¯ã€ã©ã†ã‚„ã£ã¦æ•°å¼ã‚’ãƒ—ãƒã‚°ãƒ©ãƒ ã™ã‚Œã°ã„ã„ã‹ã‚ˆãã‚ã‹ã‚‰ãªã„äººã€ã¡ã‚‡ã£ã¨ã‚„ã‚„ã“ã—ã„æ•°å¼ã«ãªã‚‹ã¨å››è‹¦å…«è‹¦ã—ã¦ã—ã¾ã„ã€ã‚³ãƒ¼ãƒ‰ã«è½ã¨ã™ã®ã«ã™ã”ãæ™‚é–“ãŒã‹ã‹ã£ã¦ã—ã¾ã†äººã€ãªã©ã€‚ ã“ã“ã§ã¯å®Ÿè¡Œé€Ÿåº¦ã«ã¤ã„ã¦ã¯ã²ã¨ã¾ãšãŠã„ã¨ã„ã¦ã€ç°¡æ½”ã§é–“é•ã„ã«ãã„ã€ã¡ã‚ƒã‚“ã¨å‹•ãã‚³ãƒ¼ãƒ‰ã‚’æ›¸ãã“ã¨ã‚’ç›®æ¨™ã«ã—ã¦ã„ã¾ã™ã€‚ Read less
nextbigthing 2013/03/21
æ©Ÿæ¢°å¦ç¿’

numpy

python
ãƒªãƒ³ã‚¯
ã‚°ãƒ¼ã‚°ãƒ«ã€ã‚ªãƒ–ã‚¸ã‚§ã‚¯ãƒˆèªè˜æŠ€è¡“ã‚’ç ”ç©¶ã™ã‚‹æ–°èˆˆä¼æ¥DNNresearchã‚’è²·åŽ
Googleã¯ã€ã‚«ãƒŠãƒ€ã«æ‹ ç‚¹ã‚’ç½®ãå¾“æ¥å“¡æ•°3åã®ç ”ç©¶ä¼æ¥DNNresearchã‚’è²·åŽã—ãŸã€‚ã“ã®ä¼æ¥ã¯éŸ³å£°èªè˜ã¨ç”»åƒèªè˜ã‚’å°‚é–€ã¨ã—ã¦ã„ã‚‹ã€‚ ã‚ªãƒ–ã‚¸ã‚§ã‚¯ãƒˆèªè˜æŠ€è¡“ã‚’å°‚é–€ã«ç ”ç©¶ã™ã‚‹ä¼æ¥ã¨ã—ã¦ã€2012å¹´ã«ãƒˆãƒãƒ³ãƒˆå¤§å¦ã®ã‚³ãƒ³ãƒ”ãƒ¥ãƒ¼ã‚¿ã‚µã‚¤ã‚¨ãƒ³ã‚¹å¦ç§‘å†…ã§ç«‹ã¡ä¸Šã’ã‚‰ã‚ŒãŸDNNresearchã¯ã€ä»Šå›žã®è²·åŽã«ã‚ˆã£ã¦Googleã®å‚˜ä¸‹ã«å…¥ã£ãŸã“ã¨ã«ãªã‚‹ã€‚ TechvibesãŒæœ€åˆã«å ±é“ã—ãŸã¨ã“ã‚ã«ã‚ˆã‚‹ã¨ã€Googleã¯ä»Šå›žã®è²·åŽã«ã‚ˆã£ã¦ã€DNNresearchã®æŠ€è¡“ã¨ã¨ã‚‚ã«ã€å‰µæ¥è€…ã§ã‚ã‚‹Geoffrey HintonåŒå¤§å¦æ•™æŽˆã®ã»ã‹ã€Alex Krizhevskyæ°ã¨Ilya Sutskeveræ°ã¨ã„ã†2åã®åŒå¤§å¦é™¢ç”Ÿã‚’ç²å¾—ã™ã‚‹ã“ã¨ã«ãªã‚‹ã¨ã„ã†ã€‚ ãƒˆãƒãƒ³ãƒˆå¤§å¦ã®David Naylorå¦é•·ã¯å£°æ˜Žã§ã€ŒGeoffrey Hintonæ•™æŽˆã®ç ”ç©¶ã¯åŸºç¤Žç ”ç©¶ã«æ ¹ã–ã—ãŸç ´å£Šçš„é©æ–°ã®ç´ æ™´ã‚‰ã—ã„ä¾‹ã§ã‚ã‚‹ã€ã¨è¿°ã¹ã‚‹ã¨ã¨ã‚‚ã«ã€Œ
nextbigthing 2013/03/13
google

deep learning

æ©Ÿæ¢°å¦ç¿’
ãƒªãƒ³ã‚¯
ãƒŠã‚¤ãƒ¼ãƒ–ãƒ™ã‚¤ã‚ºã‚’ç”¨ã„ãŸãƒ†ã‚ã‚¹ãƒˆåˆ†é¡ž - äººå·¥çŸ¥èƒ½ã«é–¢ã™ã‚‹æ–æƒ³éŒ²
ä»Šã¾ã§PRMLã‚’èªã‚“ã§å®Ÿè£…ã‚’ç¶šã‘ã¦ãã¾ã—ãŸãŒã€10ç« ã‹ã‚‰ã¯é›£ã—ãã¦æ¯ãŒç«‹ãŸãªããªã£ã¦ããŸã®ã§ã“ã“ã‚‰ã§å°‘ã—å…·ä½“çš„ãªå¿œç”¨ã«ç›®ã‚’å‘ã‘ã¦ã¿ã‚ˆã†ã¨æ€ã„ã¾ã™ã€‚æ©Ÿæ¢°å¦ç¿’ã®å¿œç”¨å…ˆã¨ã—ã¦ã¯ç”»åƒã®æ–¹ãŒçµæžœã‚’è¦‹ã¦ã„ã¦é¢ç™½ã„ã‚“ã§ã™ãŒã€å½“é¢ã¯è‡ªç„¶è¨€èªžå‡¦ç†ã‚’å–ã‚Šä¸Šã’ã¾ã™ã€‚ãã‚“ãªã‚ã‘ã§ä¸€ç•ªå§‹ã‚ã®å¿œç”¨ã¯æ©Ÿæ¢°å¦ç¿’ã¨è‡ªç„¶è¨€èªžå‡¦ç†ã®æŽ¥ç‚¹ã¨ã—ã¦éžå¸¸ã«é‡è¦ãªãƒ†ã‚ã‚¹ãƒˆåˆ†é¡žï¼ˆText Classification, Text Categorizationï¼‰ã®æŠ€æ³•ãŸã¡ã‚’è©¦ã—ã¦ã„ããŸã„ã¨æ€ã„ã¾ã™ã€‚ãƒ†ã‚ã‚¹ãƒˆåˆ†é¡žã¯æ–‡æ›¸åˆ†é¡žï¼ˆDocument Classificationï¼‰ã¨ã„ã†å‘¼ã³æ–¹ã‚‚ã‚ã‚Šã¾ã™ã€‚ãƒ†ã‚ã‚¹ãƒˆã¨æ–‡æ›¸ã¯åŒã˜æ„å‘³ã§ã™ã€‚æœ€åˆãªã®ã§è‡ªåˆ†ã®çŸ¥è˜ã®æ•´ç†ã¨å…¥é–€è€…ã¸ã®ç´¹ä»‹ã®ãŸã‚ã«ã¡ã‚‡ã£ã¨ä¸å¯§ã«ã¾ã¨ã‚ã¦ã¿ã¾ã—ãŸã€‚ ãƒ†ã‚ã‚¹ãƒˆåˆ†é¡žã¨ã¯ ãƒ†ã‚ã‚¹ãƒˆåˆ†é¡žã¨ã¯ã€ä¸Žãˆã‚‰ã‚ŒãŸæ–‡æ›¸ï¼ˆWebãƒšãƒ¼ã‚¸ã¨ã‹ï¼‰ã‚’ã‚ã‚‰ã‹ã˜ã‚ä¸Žãˆã‚‰ã‚ŒãŸã„ãã¤ã‹ã®ã‚«ãƒ†ã‚´ãƒªï¼ˆã‚¯ãƒ©ã‚¹ï¼‰ã«è‡ªå‹•åˆ†é¡žã™ã‚‹ã‚¿ã‚¹
nextbigthing 2013/02/18
æ©Ÿæ¢°å¦ç¿’

ã‚¢ãƒ«ã‚´ãƒªã‚ºãƒ

è‡ªç„¶è¨€èªžå‡¦ç†

ãƒŠã‚¤ãƒ¼ãƒ–ãƒ™ã‚¤ã‚º
ãƒªãƒ³ã‚¯
PRML/course - æ©Ÿæ¢°å¦ç¿’ã®ã€Œæœ±é·ºã®æœWikiã€
å¿…è¦ãªäº‹å‰çŸ¥è˜â€ å¾®ç©åˆ†ï¼šå¤šå¤‰é‡ã®å¾®åˆ†ã‚„ç©åˆ†ã¯å¿…è¦ã§ã™ï¼Žãƒ©ã‚°ãƒ©ãƒ³ã‚¸ãƒ¥ã®æœªå®šä¹—æ•°æ³•ã‚„ç°¡å˜ãªå¤‰åˆ†ã‚‚å¿…è¦ã§ã™ãŒã“ã‚Œã‚‰ã¯ä»˜éŒ²ã«ç°¡å˜ãªè§£èª¬ãŒã‚ã‚Šã¾ã™ï¼Žè¤‡ç´ æ•°ã®å¾®ç©åˆ†ã«ã¤ã„ã¦ã¯ä¸è¦ã§ã™ï¼Ž ç·šå½¢ä»£æ•°ï¼š3Ã—3ä»¥ä¸Šã®å›ºæœ‰å€¤ãƒ»é€†è¡Œåˆ—ï¼Œè¡Œåˆ—ã«å¯¾ã™ã‚‹å¾®ç©åˆ†ã¯å¿…è¦ã§ã™ãŒï¼Œã“ã‚Œã‚‰ã«ã¤ã„ã¦ã¯ä»˜éŒ²ã«ç°¡å˜ã«ã¾ã¨ã‚ã‚‰ã‚Œã¦ã„ã¾ã™ï¼Ž$l^2$ç©ºé–“ã‚„ãƒ’ãƒ«ãƒ™ãƒ«ãƒˆç©ºé–“ï¼Œè¤‡ç´ æ•°ã«ã¤ã„ã¦ã¯ä¸è¦ã§ã™ï¼Ž ç¢ºçŽ‡ï¼šåŸºç¤Žã«ã¤ã„ã¦ã¯1ç« ã§è¿°ã¹ã‚‰ã‚Œã¦ã„ã¾ã™ãŒï¼Œç¢ºçŽ‡ã®æ‰±ã„ã‚„è€ƒãˆã«æ…£ã‚Œã¦ã„ã‚‹ã¨ä¾¿åˆ©ã§ã™ï¼Žç¢ºçŽ‡æ¸¬åº¦ãªã©å…¬ç†ã«ã¾ã§è¸ã¿è¾¼ã‚“ã çŸ¥è˜ã‚„ï¼Œçµ±è¨ˆã®æ¤œå®šãªã©ã®çŸ¥è˜ã¯ä¸è¦ã§ã™ï¼Ž â†‘ æœ¬æ›¸ã«é–¢é€£ã™ã‚‹åˆ†é‡Žâ€ ä»¥ä¸‹ã®åˆ†é‡Žã§ãƒ‡ãƒ¼ã‚¿ã‹ã‚‰ã®äºˆæ¸¬ã‚„åˆ†æžãŒå¿…è¦ã¨ã•ã‚Œã‚‹åˆ†é‡Ž å¯¾è±¡åˆ†é‡Žï¼šæœ¬æ›¸ã¨ç›´æŽ¥çš„ãªé–¢é€£ãŒã‚ã‚‹åˆ†é‡Žã«ã¯æ¬¡ã®ã‚ˆã†ãªã‚‚ã®ãŒã‚ã‚Šã¾ã™ï¼Ž æ©Ÿæ¢°å¦ç¿’ï¼Œãƒ‘ã‚¿ãƒ¼ãƒ³èªè˜ï¼Œçµ±è¨ˆï¼Œãƒ‡ãƒ¼ã‚¿ãƒžã‚¤ãƒ‹ãƒ³ã‚° å¿œç”¨åˆ†é‡Žï¼šæ©Ÿæ¢°å¦ç¿’ï¼Œãƒ‘ã‚¿ãƒ¼ãƒ³èªè˜ï¼Œçµ±è¨ˆçš„äºˆæ¸¬æŠ€è¡“ãŒå¿œç”¨ã•ã‚Œã¦ã„ã‚‹åˆ†é‡Žã«ã¯æ¬¡ã®ã‚ˆã†ãªã‚‚
nextbigthing 2013/02/08
æ©Ÿæ¢°å¦ç¿’

PRML
ãƒªãƒ³ã‚¯
ã¯ã¦ãªãƒ–ãƒã‚° | ç„¡æ–™ãƒ–ãƒã‚°ã‚’ä½œæˆã—ã‚ˆã†
12æœˆ3æ—¥éŸ“å›½ã«æˆ’åŽ³ä»¤ãŒå®£è¨€ã•ã‚ŒãŸã€‚ãã—ã¦å¯’ç©ºã®å…ƒã«å±Šã‘ã‚‰ã‚Œã‚‹ã‚³ãƒ¼ãƒ’ãƒ¼ã€‚ 12æœˆ3æ—¥22æ™‚åŠã”ã‚ã€ ã©ã†ã„ã†ã‚ã‘ã‹å°¹éŒ«æ‚¦å¤§çµ±é ˜ã«ã‚ˆã‚Šéžå¸¸æˆ’åŽ³ãŒå®£è¨€ã•ã‚ŒãŸã€‚ãã‚Œã«ã¤ã„ã¦ã€éŸ“å›½ã§ç”Ÿæ´»ã‚’ã™ã‚‹ç§ã®è¦–ç·šã§è¨˜éŒ²ã‚’ã—ã¦ãŠããŸã„ã¨æ€ã†ã€‚ä»Šé€±ã¯æ›¸ãã‹ã‘ã ã£ãŸé«˜é›„&å°å—æ—…è¡Œè¨˜ã‚’ã‚¹ã‚¿ãƒ¼ãƒˆã•ã›ãŸã‹ã£ãŸãŒã€å¾Œå›žã—ã«ã™ã‚‹ã€‚ ï¼ˆå†™çœŸã¯ä»Šå›žã®å¼¾åŠ¾ã‚’æ±‚ã‚ã‚‹ãƒ‡ãƒ¢ã§â€¦
nextbigthing 2012/12/21
python

ãƒ‘ãƒ¼ã‚»ãƒ—ãƒˆãƒãƒ³

æ©Ÿæ¢°å¦ç¿’
ãƒªãƒ³ã‚¯
ãƒ‹ãƒ¥ãƒ¼ãƒ©ãƒ«ãƒãƒƒãƒˆã®é€†è¥² - Preferred Networks Research & Development
å²¡é‡ŽåŽŸã§ã™ã€‚Deep LearningãŒå„åˆ†é‡Žã®ã‚³ãƒ³ãƒšãƒ†ã‚£ã‚·ãƒ§ãƒ³ã§å„ªå‹ã—è©±é¡Œã«ãªã£ã¦ã„ã¾ã™ã€‚Deep Learningã¯7ã€8æ®µã¨æ·±ã„ãƒ‹ãƒ¥ãƒ¼ãƒ©ãƒ«ãƒãƒƒãƒˆã‚’ä½¿ã†å¦ç¿’æ‰‹æ³•ã§ã™ã€‚ã™ã§ã«ã€ç”»åƒèªè˜ã€éŸ³å£°èªè˜ã€æœ€ã‚‚æœ€è¿‘ã§ã¯åŒ–åˆç‰©ã®æ´»æ€§äºˆæ¸¬ã§å„ªå‹ã—ãŸã‚Šã€æ—¢å˜ãƒ‡ãƒ¼ã‚¿ãƒ»ã‚»ãƒƒãƒˆã§ã®æœ€é«˜ç²¾åº¦ã‚’é”æˆã—ã¦ã„ã¾ã™ã€‚ä»¥ä¸‹ã«å¹¾ã¤ã‹ä¾‹ã‚’ã‚ã’ã¾ã™ã€‚ ç”»åƒèªè˜ LSVRC 2012 [html] Â å„ªå‹ãƒãƒ¼ãƒ ã‚¹ãƒ©ã‚¤ãƒ‰ [pdf], ã¾ã¨ã‚ã‚¹ãƒ©ã‚¤ãƒ‰[pdf] Googleã«ã‚ˆã‚‹å·¨å¤§ãªNeuralNetã‚’åˆ©ç”¨ã—ãŸç”»åƒèªè˜ï¼ˆçŒ«èªè˜ã¨ã—ã¦æœ‰åï¼‰[paper][slide][æ—¥æœ¬èªžè§£èª¬] ã¾ãŸã€å„åˆ†é‡Žã®ãƒˆãƒƒãƒ—ã‚«ãƒ³ãƒ•ã‚¡ãƒ¬ãƒ³ã‚¹ã§Deep Learningã®ãƒãƒ¥ãƒ¼ãƒˆãƒªã‚¢ãƒ«ãŒè¡Œã‚ã‚Œã€ã‚µãƒ¼ãƒ™ã‚¤è«–æ–‡ã‚‚ã„ãã¤ã‹å‡ºã¾ã—ãŸã€‚ãŠãã‚‰ãæ¥å¹´ä»¥é™ã“ã†ã—ãŸè©±ãŒå¢—ãˆã¦ãã‚‹ã“ã¨ãŒè€ƒãˆã‚‰ã‚Œã¾ã™ã€‚ ICML 2012 [pdf] ACL 2012 [pdf] CVPR
nextbigthing 2012/11/02
äººå·¥çŸ¥èƒ½

è«–æ–‡

æ©Ÿæ¢°å¦ç¿’

ãƒ‡ãƒ¼ã‚¿ãƒžã‚¤ãƒ‹ãƒ³ã‚°

deep learning
ãƒªãƒ³ã‚¯
é€Ÿæ°´æ¡ƒåã€Œãƒ‘ã‚¿ãƒ¼ãƒ³èªè˜ã¨æ©Ÿæ¢°å¦ç¿’å…¥é–€ã€
250. Reference "Pattern Recognition and Machine Learning" Christopher M. Bishop Springer; 1st ed. 2006. Corr. 2nd printing edition (October 1, 2007) "Truth and Probability" Frank Plumpton Ramsey (1926) "The physical basis of IMRT and inverse planning" S Webb British Journal of Radiology (2003) 76, 678-689 251. Wikipedia æ¸¡è¾ºæ…§ http://ja.wikipedia.org/wiki/%E6%B8%A1%E8%BE%BA%E6%85 %A7 ã€ŽNo Free Lunch T
nextbigthing 2012/10/20
æ©Ÿæ¢°å¦ç¿’

ãƒ‡ãƒ¼ã‚¿ãƒžã‚¤ãƒ‹ãƒ³ã‚°
ãƒªãƒ³ã‚¯
çµ±è¨ˆçš„æ©Ÿæ¢°å¦ç¿’ã«ãŠã‘ã‚‹é‡åã‚¢ãƒ‹ãƒ¼ãƒªãƒ³ã‚°
KDD2014å‹‰å¼·ä¼š: Large-Scale High-Precision Topic Modeling on Twitter
nextbigthing 2012/10/20
ãƒ‡ãƒ¼ã‚¿ãƒžã‚¤ãƒ‹ãƒ³ã‚°

æ©Ÿæ¢°å¦ç¿’
ãƒªãƒ³ã‚¯
çµ±è¨ˆçš„æ©Ÿæ¢°å¦ç¿’å…¥é–€
çµ±è¨ˆçš„æ©Ÿæ¢°å¦ç¿’å…¥é–€(under construction) æ©Ÿæ¢°å¦ç¿’ã®æ´å²ppt pdf æ´å²ä»¥å‰ äººå·¥çŸ¥èƒ½ã®æ™‚ä»£ å®Ÿç”¨åŒ–ã®æ™‚ä»£ å°Žå…¥ppt pdf æƒ…å ±ã®å¤‰æ›éŽç¨‹ã®ãƒ¢ãƒ‡ãƒ«åŒ– ãƒ™ã‚¤ã‚ºçµ±è¨ˆã®æ„ç¾© è˜åˆ¥ãƒ¢ãƒ‡ãƒ«ã¨ç”Ÿæˆãƒ¢ãƒ‡ãƒ« æ¬¡å…ƒã®å‘ªã„ æå¤±é–¢æ•°, bias, variance, noise ãƒ‡ãƒ¼ã‚¿ã®æ€§è³ª æ•°å¦ã®ãŠã•ã‚‰ã„ppt pdf ç·šå½¢ä»£æ•°å¦ã§å½¹ç«‹ã¤å…¬å¼ ç¢ºçŽ‡åˆ†å¸ƒ æƒ…å ±ç†è«–ã®è«¸æ¦‚å¿µ (KL-divergenceãªã©) ç·šå½¢å›žå¸°ã¨è˜åˆ¥ppt pdf ç·šå½¢å›žå¸° æ£è¦æ–¹ç¨‹å¼ æ£è¦åŒ–é …ã®å°Žå…¥ ç·šå½¢è˜åˆ¥ ãƒ‘ãƒ¼ã‚»ãƒ—ãƒˆãƒãƒ³ ã‚«ãƒ¼ãƒãƒ«æ³•ppt pdf ç·šå½¢è˜åˆ¥ã®ä¸€èˆ¬åŒ– ã‚«ãƒ¼ãƒãƒ«ã®æ§‹ç¯‰æ³• æœ€å¤§ãƒžãƒ¼ã‚¸ãƒ³åˆ†é¡žå™¨ ã‚½ãƒ•ãƒˆãƒžãƒ¼ã‚¸ãƒ³ã®åˆ†é¡žå™¨ SVMã«ã‚ˆã‚‹å›žå¸°ãƒ¢ãƒ‡ãƒ« SVMå®Ÿè£…ä¸Šã®å·¥å¤« ã‚¯ãƒ©ã‚¹ã‚¿ãƒªãƒ³ã‚°ppt pdf è·é›¢ã®å®šç¾© éšŽå±¤åž‹ã‚¯ãƒ©ã‚¹ã‚¿ãƒªãƒ³ã‚° K-means ãƒ¢ãƒ‡ãƒ«æŽ¨å®šppt pdf æ½œåœ¨å¤‰æ•°ã®ã‚ã‚‹ãƒ¢ãƒ‡ãƒ« EMã‚¢ãƒ«
nextbigthing 2012/10/18
æ©Ÿæ¢°å¦ç¿’

ãƒ‡ãƒ¼ã‚¿ãƒžã‚¤ãƒ‹ãƒ³ã‚°

ã‚¢ãƒ«ã‚´ãƒªã‚ºãƒ

çµ±è¨ˆ
ãƒªãƒ³ã‚¯
Pythonã§PLSAã‚’å®Ÿè£…ã—ã¦ã¿ã‚‹
probabilistic latent semantic analysis (PLSA)ã¯ã€ ãƒ»æ–‡æ›¸dãŒP(d)ã§é¸ã°ã‚Œã‚‹ ãƒ»æ½œåœ¨å¤‰æ•°zãŒP(z|d)ã§é¸ã°ã‚Œã‚‹ ãƒ»èªžwãŒP(w|z)ã§ç”Ÿæˆã•ã‚Œã‚‹ ã¨ã„ã†ãƒ—ãƒã‚»ã‚¹ã‚’çµŒã¦ã€çµæžœã¨ã—ã¦(d,w)ã®ãƒšã‚¢ãŒè¦³æ¸¬ã•ã‚Œã‚‹ã¨ã„ã†æ–‡æ›¸ã¨èªžã®ç”Ÿæˆãƒ¢ãƒ‡ãƒ«ã€‚ å¼ã§è¡¨ã™ã¨ (1) ã¨ãªã‚‹ã€‚P(d,w)ã®å°¤ã‚‚ã‚‰ã—ã„ç¢ºçŽ‡åˆ†å¸ƒã‚’è¦‹ã¤ã‘ãŸã„ã€‚å¯¾æ•°å°¤åº¦é–¢æ•°ã¯ (2) ã¨ãªã‚‹ã€‚n(d,w)ã¯èªžwãŒæ–‡æ›¸dã«å‡ºç¾ã™ã‚‹å›žæ•°ã€‚ã“ã®å¼ã¯è¨“ç·´ãƒ‡ãƒ¼ã‚¿n(d,w)ï¼ˆ;ã©ã®èªžãŒã©ã®æ–‡æ›¸ã«ä½•å›žå‡ºç¾ã—ãŸã‹ï¼‰ãŒå°¤ã‚‚ã‚‰ã—ã„ç¢ºçŽ‡åˆ†å¸ƒP(d,w)ã«å¾“ã†ã¨ãæœ€å¤§ã«ãªã‚‹ã€‚ãƒ™ã‚¤ã‚ºã®å®šç†ã‚’ç”¨ã„ã‚‹ã¨ (3) ã¨ãªã‚‹ã“ã¨ã‚’åˆ©ç”¨ã—ã¦ã€ã“ã®å°¤åº¦é–¢æ•°ã‚’æœ€å¤§åŒ–ã™ã‚‹ãŸã‚ã«EMã‚¢ãƒ«ã‚´ãƒªã‚ºãƒ ã‚’ç”¨ã„ã¦å®Ÿè£…ã—ã¦ã¿ã‚‹ã€‚ï¼ˆéŽå¦ç¿’ã‚’å›žé¿ã™ã‚‹ãŸã‚ã«æ–‡çŒ®ã§ã¯Tempered EM (TEM)ã‚’ç”¨ã„ã¦ã„ã‚‹ã€‚ï¼‰å°¤åº¦é–¢æ•°ãŒåŽæŸã™ã‚‹ã¾ã§ä»¥ä¸‹ã®E-ste
nextbigthing 2012/10/16
python

plsi

æ©Ÿæ¢°å¦ç¿’

ã‚¢ãƒ«ã‚´ãƒªã‚ºãƒ
ãƒªãƒ³ã‚¯
ã‚µãƒ¼ãƒ“ã‚¹çµ‚äº†ã®ãŠçŸ¥ã‚‰ã› - NAVER ã¾ã¨ã‚
ã‚µãƒ¼ãƒ“ã‚¹çµ‚äº†ã®ãŠçŸ¥ã‚‰ã› NAVERã¾ã¨ã‚ã¯2020å¹´9æœˆ30æ—¥ã‚’ã‚‚ã¡ã¾ã—ã¦ã‚µãƒ¼ãƒ“ã‚¹çµ‚äº†ã„ãŸã—ã¾ã—ãŸã€‚ ç´„11å¹´é–“ã€NAVERã¾ã¨ã‚ã‚’ã”åˆ©ç”¨ãƒ»ã”æ„›é¡§ã„ãŸã ãèª ã«ã‚ã‚ŠãŒã¨ã†ã”ã–ã„ã¾ã—ãŸã€‚
nextbigthing 2012/09/10
æ©Ÿæ¢°å¦ç¿’

çµ±è¨ˆ
ãƒªãƒ³ã‚¯
äººå·¥çŸ¥èƒ½å¦ä¼šèªŒã€Œç§ã®ãƒ–ãƒƒã‚¯ãƒžãƒ¼ã‚¯ã€
äººå·¥çŸ¥èƒ½å¦ä¼šèªŒã§ã¯ï¼Œäººå·¥çŸ¥èƒ½ç ”ç©¶ã®å„åˆ†é‡Žã®å°‚é–€å®¶ãŒWebãƒšãƒ¼ã‚¸ã‚’ç´¹ä»‹ã™ã‚‹ã€Œç§ã®ãƒ–ãƒƒã‚¯ãƒžãƒ¼ã‚¯ã€ã‚’é€£è¼‰ä¸ã§ã™ï¼Ž ã“ã“ã«ã¯ï¼Œã“ã®ã€Œç§ã®ãƒ–ãƒƒã‚¯ãƒžãƒ¼ã‚¯ã€ã«æŽ²è¼‰ã•ã‚ŒãŸãƒšãƒ¼ã‚¸ã¸ã®ãƒªãƒ³ã‚¯ã‚’é›†ã‚ã¾ã—ãŸï¼Ž ã¿ãªã•ã¾ã®ç ”ç©¶ã«ã”æ´»ç”¨ãã ã•ã„ï¼Ž Vol.23
nextbigthing 2012/08/24
æ©Ÿæ¢°å¦ç¿’

äººå·¥çŸ¥èƒ½

ç ”ç©¶
ãƒªãƒ³ã‚¯
ãƒ‘ãƒ¼ã‚»ãƒ—ãƒˆãƒãƒ³ã‚’Pythonã§ã‹ã„ã¦ã¿ãŸï¼Ž - Screaming Loud
æ©Ÿæ¢°å¦ç¿’ã®åˆæ©ã¨ã„ã†ã“ã¨ã§ãƒ‘ãƒ¼ã‚»ãƒ—ãƒˆãƒãƒ³ã‚’å®Ÿè£…ã—ã¦ã¿ã¾ã—ãŸï¼Ž ãƒ‘ãƒ¼ã‚»ãƒ—ãƒˆãƒãƒ³ã¨ã¯ï¼’å€¤è˜åˆ¥å™¨ã§ã™ï¼Ž ã‚¯ãƒ©ã‚¹ã¯ï¼‘ã‚¤ãƒ†ãƒ¬ãƒ¼ã‚·ãƒ§ãƒ³ã‚’è¡¨ã—ã¦ã„ã¾ã™ï¼Ž 2æ¬¡å…ƒä»¥ä¸Šã§ã‚‚è¨ˆç®—ã§ãã‚‹ã‚ˆã†ã«ã—ã¾ã—ãŸï¼Ž import sys from optparse import OptionParser import matplotlib as mpl mpl.use('PDF') import matplotlib.pyplot as plt import numpy import random def get_train_data(filename): '''make dictionary from training data''' train = [] with open(filename) as f: for line in f: line = line.split() train.append((line)) r
nextbigthing 2012/08/07
æ©Ÿæ¢°å¦ç¿’

ãƒ‘ãƒ¼ã‚»ãƒ—ãƒˆãƒãƒ³
ãƒªãƒ³ã‚¯
FrontPage - æƒ…å ±è«–çš„å¦ç¿’ç†è«–ã¨æ©Ÿæ¢°å¦ç¿’ã®ã€Œæœ±é·ºã®æœWikiã€
æœ±é·ºã®æœWikiï¼ˆã¨ãã®ã‚‚ã‚Š ã†ãƒãï¼‰â€ æœ±é·ºã®æœWikiã¯ï¼Œæ©Ÿæ¢°å¦ç¿’ã«é–¢é€£ã—ãŸï¼Œãƒ‡ãƒ¼ã‚¿ãƒžã‚¤ãƒ‹ãƒ³ã‚°ï¼Œæƒ…å ±ç†è«–ï¼Œè¨ˆç®—è«–çš„å¦ç¿’ç†è«–ï¼Œçµ±è¨ˆï¼Œçµ±è¨ˆç‰©ç†ã«ã¤ã„ã¦ã®æƒ…å ±äº¤æ›ã®å ´ã§ã™ï¼Žã“ã‚Œã‚‰æ©Ÿæ¢°å¦ç¿’é–¢ä¿‚ã®è©±é¡Œï¼Œãƒªãƒ³ã‚¯ï¼Œé–¢é€£äº‹é …ï¼Œæ›¸ç±ãƒ»è«–æ–‡ç´¹ä»‹ãªã©ã®æƒ…å ±ã‚’æ‰±ã„ã¾ã™ï¼Ž æ›´æ–°ã•ã‚ŒãŸãƒšãƒ¼ã‚¸ã‚’ç¢ºèªã™ã‚‹ã«ã¯RSSãƒªãƒ¼ãƒ€ã‚’ä½¿ã£ã¦å³ä¸‹ã®RSSãƒªãƒ³ã‚¯ã‚’ãƒã‚§ãƒƒã‚¯ã™ã‚‹ã‹ï¼Œæœ€çµ‚æ›´æ–°ã®ãƒšãƒ¼ã‚¸ã‚’å‚ç…§ã—ã¦ãã ã•ã„ï¼Ž ãƒšãƒ¼ã‚¸ã®ä¸ã§ã©ã“ãŒæ›´æ–°ã•ã‚ŒãŸã‹ã‚’è¦‹ã‚‹ã«ã¯ï¼Œä¸Šã®ã€Œå·®åˆ†ã€ã‚’ã‚¯ãƒªãƒƒã‚¯ã—ã¦ä¸‹ã•ã„ï¼Ž æ•°å¼ã®è¡¨ç¤ºã« MathJax ã‚’åˆ©ç”¨ã—ã¦ã„ã¾ã™ï¼Žæ•°å¼ã®ä¸Šã§ã‚³ãƒ³ãƒ†ã‚ã‚¹ãƒˆãƒ¡ãƒ‹ãƒ¥ãƒ¼ã‚’ä½¿ã†ã¨å„ç¨®ã®è¨å®šãŒå¯èƒ½ã§ã™ï¼Žç‰¹ã«è¨å®šã‚’ã—ãªãã¦ã‚‚æ•°å¼ã¯é–²è¦§ã§ãã¾ã™ãŒï¼Œãƒ•ã‚©ãƒ³ãƒˆã‚’ã‚¤ãƒ³ã‚¹ãƒˆãƒ¼ãƒ«ã™ã‚Œã°æ•°å¼ã®è¡¨ç¤ºãŒãã‚Œã„ã§é«˜é€Ÿã«ãªã‚Šã¾ã™ï¼Žè©³ã—ãã¯ æ•°å¼ã®è¡¨ç¤º ã®ãƒšãƒ¼ã‚¸ã‚’å‚ç…§ã—ã¦ä¸‹ã•ã„ï¼Ž ã”ãç°¡å˜ãªWikiã®ä½¿ã„æ–¹ãŒã“ã®ãƒšãƒ¼ã‚¸ã®æœ€å¾Œã«ã‚ã‚Šã¾ã™ï¼Žãƒˆãƒƒãƒ—ãƒšãƒ¼ã‚¸ã‚„ãƒ¡ãƒ‹ãƒ¥ãƒ¼ãªã©
nextbigthing 2012/08/07
æ©Ÿæ¢°å¦ç¿’

ãƒ‡ãƒ¼ã‚¿ãƒžã‚¤ãƒ‹ãƒ³ã‚°
ãƒªãƒ³ã‚¯
1 2 æ¬¡ã®ãƒšãƒ¼ã‚¸