2003-07-01ãã1ã¶æéã®è¨äºä¸è¦§
ãã ã®GIFã¢ãã¡ãç«ä½ã«è¦ãããç«ä½ã£ã¦ãããããã¨ã ã£ããã§ããç¥æ§ã é¢ç½ãã®ã¯ãæåã¯ã´ãã´ãåãã¦ããã ãã§ããã¡ï¼ãã£ã¦æããªã®ã«ãæ £ããã¨ããç«ä½ã«ããè¦ããªããªãã¨ããï¼ã£ã¦ãåã ãã§ããããï¼ãæãå¹æçãªåããå¨æã£ã¦ããâ¦
ç¶ç·¨ãããã®ãµã¤ãã®ä»ã®èªã¿ç©ãã¨ã¦ãé¢ç½ããwww.textfile.orgããã ãä¸ã¤æ°ãããââæ°ã§ãã£ã¦æµè¡ããªãããªã ä¸ã¤æ°ããããã£ããããæ°ã§ ã¨ãã ä¸ã¤æ°ããã4æ¡ã®ç´ æ°ã§ ã¨ãã
æãã¤ãã¤ãã§ã«ããã²ã¨ã¤ã å½¢æ ç´ è§£æã§ããã¯ã«ãªãã®ãæªç¥èªã®å¦çã ãã©ããã¡ãã¡è¾æ¸ç»é²ãã¦ã³ã³ãã¤ã«ãã¦ãã£ã¦ã®ã¯çµæ§é¢åããªã®ã§ãã¿ããªã§ãã£ã¦ããã£ã¦ç»é²ã§ããä»çµã¿ãä½ã£ããé¢ç½ããã ãè¾æ¸ã ãå ±æãã¦ãé¢ç½ã¿ããªãã®ã§ããã£â¦
ååããã®ã¨ããããã ã¡ãã£ã¨æåãé³æ¥½ã®ä¿®è¡ã«ãéãããã®ãããã¾ããªã
ç¾å æ¬èããã®ããï¼ä¸ã«äºãã®ç¨®ã¯å°½ãã¾ãï¼ãèªãã§ã®æãã¤ãã ãã®ãã¼ã¯ã¼ã群ãç¡è³ã®è¾æ¸ã«ä½¿ããªãããªã解説æä¸ã®ãã¼ã¯ã¼ãã®ãªã³ã¯ã¨ããââãå«ããã¼ã¯ã¼ããã®ãªã³ã¯ããã©ã£ãããããã¨ã§ããç¨åº¦é層åã§ãããã ãããããã¯ãããªæãâ¦
ãã¡ã ãããã¡ãããã¾ãé«å§çã¨ãè¨ã人ãåºã¦ãããï¼ç¬ï¼ãªã®ã§ãåã ã£ããã©ããããã¡ãã£ã¨èãã¦ã¿ãã ã§ãä»ã®ã¨ããã®çµè«ãããåãè¿ãã°è¢å°è·¯ã«ãªããªãããããªãããªã ãââããã¦ã¯ãããªããçç±ãè«ççã«èª¬æã§ããªãå ´åãããã¯èªâ¦
ã¯ã¦ãªã®è³ªåãåºã¾ãããââã¯ä½æ ãããªãã®ãããââã«ã¯æ´å©äº¤éã®ä»ãå®çªã®ã人ã殺ããã¨ãããè°ä¼ã«ãã¹ã¯ãã¤ããã¾ã¾å ¥ããã¨ããªã©ãå ¥ãã¾ãã åã¯ãã®æã®è³ªåã大å«ããªãã ããªãã質åãã人ã¯ç¡éªæ°ã«èãã¦ãå ´åãããããï¼ç¹ã«åä¾ã®å ´â¦
essaããã®ãã®è¨äºãèªãã§æã£ããã¨ããªã³ã¯å ãã¤ãèªã¿ãµãã£ã¦ãã¾ã£ãã åãæãã«ãæ¬å½ã«å¹¸ãã«ãªãã³ãã¨ã¯ãèªåèªèº«ã®ä¾¡å¤ãä»äººã®è©ä¾¡ã§æ¸¬ããªããã¨ã ã人ãèªåãã©ãæã£ã¦ãããã§ä¸åä¸æããã®ã馬鹿ãããããæã£ãã¨ããè©ä¾¡ãã¦ããâ¦
é¢ç½ãããªã®ã§ã¡ã¢ãå¾ã§èªãã§ã¿ãããã¡ãªã¿ã«ãã«ã¿ã¤ãã«ã¯ Towards Answering Opinion Questons: Separating Facts from Opinions and Identifying the Polarity of Opinion Sentences é·ããã ãã©ãããã®ãã¾ããæ¥è¨ããã
é è·é¢ã®æ¹ã試ãã¦ã¿ã¦ã¯å¦ä½ãåã¯å¼ã£è¶ãåã ã£ãã6ã¶æã§5ååè¿ããããã®ã«ãªãããã¾ãè¿ãã¨ãã¡ã¿ããã§ãã ãã©ãããã®ãã¾ããæ¥è¨ããã
ããã§æ°ä»ãããã°ã®ä¿®æ£ãä¸å¿ãé£æ³ã®å®è£ ã®ã¤ãããªãã ãã©ããããã¨ãã¡ãããã«ãªã£ããããã¾ãããç¡è³ã ãããã¯ã¯ã
çµåããã®www.textfile.orgãããã½ã¯ã©ãã¹çã§ããªã åããããã®ããæ¹è©¦ããã ãã©ããã¾ã«ã²ã©ã失æããããéä¸ã§ãã¡ãã®æå³ãç¸æã«ä¼ãã£ã¦ãã¾ããçµå±è¨ãéãããã¦ãã¾ãã®ã ã ãã¡ããããè¨ã£ããç¸æã¯ããè¨ãã ããããããããããâ¦
森山ããã®ã¨ããããããããã¾ãç©æ¬²ãã»ã»ã»ããããå·éã«çºãã¦ã¿ãã¨ã1/4ãããã¯æ¢ã«æã£ã¦ããæãã追è¨ï¼22åã§ãããã 寺ç°å¯ 彦ãæåã«æãããã¦ãããããã好ãã ãªãã
使ãæ¹ã¨ä¼è©±ä¾ã追å ã ã¾ããå®ã¯å æ¥ircåå ã®å®é¨ããã¦ã¿ãã¨ããããã£ããæåãã¦ãã¾ã£ããå¼è¶ãã®çä»ããé²ãã§æã®PCããã³ãã¼ã«ã®å±±ããçºæãããããIRC #Margsã¸å¸¸é§ãããããã«ããã£ã¨ã
Next stepãæ¢ãããã«ï¼ç¬ï¼è²ã è«æãèªãã§ãåå¼·ä¸ã 森ããã®è«æãåããæ°ã«ãªã£ã¦ããã®ã§ãããæ©ä¼ã ããæ¾ãèªã¿ãã¦ããã ãã©ããå½¢æ ç´ bi-gramã¨åè©bi-gramã®éãåããã«ããå½¢æ ç´ è§£æ(303[KB], PSãã¡ã¤ã«)ããèªãã§ãããã§ææ¡ããã¦ãâ¦
ããªã楽ããããã¯ã次ã®èª²é¡ã¯ãGhotiãããã¸åå ããããã¨ã ãªãã
ãã¼ã¯ã¼ãã®å¦çã¾ããã§ãã°çºè¦ãã¦ããããã°ã¨ããããã¯å®è£ å¿ãã å¦çã®æµãã¨ãã¦ã 人éã®çºè¨ããããã¤ããã¼ã¯ã¼ããåãåºã ãã®è¤æ°ã®ãã¼ã¯ã¼ãããã©ã³ãã ã«1ã¤åãåºã㦠ãã®ãã¼ã¯ã¼ããå ã«é¢é£èªãé¸ã¶ ã¯ãã ã£ããã ãã©ãæå¾ã®â¦
é¨å®¤ã§çµ¶è³ã®åµã確ãã«ãã£ã¨æ¥ããã®ãããã山形ãããããããç´ ç´ãªãç´çã®æç« ããã¾ããªãã Amazonã§ã½ã¡ã£ã¨è²·ãããã®ããã£ã¨ãããã¦ãåã«ãèªã¿ããªãããã®ãã©ãããã¾ãã¯æ¸åºã§ç¢ºèªããããã§ãã¿ããã£ã¦ããã¾ãç½®ãã¦ãæ¬å±ãªãããçµâ¦
hakunamatataããã®æ¥è¨ãããæå¾ã®æ³åï¼ããªã¿ã¯ã«ã¯ããããããã - ããããã¯å½¼ãã®ãã¨ã§åããã¨ã«ãªããããããªããããï¼ãé¤ãã°ãããã¨ããããã³ãã¢åãï¼ã¨ãéãããï¼ã®æè¨ã§ããå人çã«ã¯æ¿ããåæã ãããã³ãã¢ãã³ãã¢ããæ以â¦
ãã¡ã¤ã«ç½®ãå ´ã«ã¢ãããã¼ããã¾ãããrubyã¨MeCabãå¿ è¦ã§ãã 使ãæ¹ã¯ããããã®sixamoã¨ã»ã¼åãã§ãï¼é常ã«åèã«ããã¦ããã ãã¾ãããæè¬ãã¾ãï¼ããã¼ã¿ç½®ãå ´ã®ãã£ã¬ã¯ããªã«ghoti.datã¨è¨ããã¡ã¤ã«ãããã¨ããããå¦ç¿ãã¡ã¤ã«ã¨ã¿ãªããâ¦
å½ç«ææçç 究æ ææçæ å ±ã»ã³ã¿ã¼ãä¸ä½ãããèãã¦ããã®ã§ãããããé¢ç½ãã®ã§ç¡æã§ãªã³ã¯ãã¡ãããããã¯ã¯ã 森山ããã®æ¥è¨ããã
ã¿ã¤ãã«ã¯å¤§ä»°ã ãã©ããå§ãã 3é¨æ§æã§ãªã£ã¦ãã¦ã第1é¨ã¯CFOã¨è¨ããã¸ã·ã§ã³ã確ç«ããçµç·¯ã®ã¾ã¨ãã第2é¨ã¯ãã¹ã¼ãã¼CFOãã®ã¾ãã«åä¼ã第3é¨ã¯CFOæ©è½ï¼æ¥æ¬ä¼æ¥ã«æ¬ ãã¦ãããã®ãã¨ã®ä½ç½®ä»ããããä»å¾æ¥æ¬ä¼æ¥ãä½ã«ãã£ã¬ã³ã¸ãã¦ããã°è¯â¦
ã¡ã¢ãçµåããã®www.textfile.orgãããæ°å±±ããã®è¨³ã¨ãã®åæãã
ã¢ã³ããããå·¡åããã ãã§åæ¥æ½°ããããæ´ã«æã£ããã¨/èãããã¨ãã¡ã¢ãããã¨ããããããã1æ¥æ½°ãããæ´ã«æ´ã«ãå·¡åå ã§é¢ç½ããªã³ã¯ãæ²ããã«ã¢ã³ããã«è¿½å ãããããã®ã§ããã®ããã»ã¹ã¯ã©ãã©ãæéããããããã«ãªãã®ã ãããã»ã©ã»ã©ã«â¦
æ°å± ã¯æ³åãã¦ãã以ä¸ã«å¿«é©ã西è»ã°ã¬ã¼ãããã ãããçä»ãã¯ã¾ã ã¾ã ç¶ãã®ã§ãã£ãããªããåå¹´ãããçä»ããªãããã»ã»ã»ã
ãªãããã¯ã¦ãªã®æ¥è¨ã£ã¦Googleã®ã©ã³ã¯è¯ãããããªãã§ãããããGraphviz cygwinãã§ãã®æ¥è¨ããããã«åºã¦ãã¡ããã®ã¯æµç³ã«ã¡ãã£ã¨ã¤ã«ã¬ãªã¢ãã«ã¨æããã¾ããã¾ããããã ãã©ã ã§ãåãGoogleã®äººéã ã£ãããã¯ã¦ãªã®æ¥è¨ç¾¤å¯¾çããã¨æããªã
å°åå対çã®ä¸ç°ãªãã§ãã¨ãããæ°ãããã§ãããã ãã¡ãããä¸å¦æ²»çã«å©æéãåºããã¨èªä½ã¯è¯ããã¨ã ã¨æãã¾ãããã©ããããæ¬å½ã«å°åå対çã«ãªããã©ããã¯ãããã¦çåãä¸å¦æ²»çã«ãéãããããã人å£ãæ¸ã£ã¦ããã§ããï¼ãããããªãã§ãâ¦
ããã®è¨äºã§ç¥ãã¾ããã ãã®è¨äºï¼FTãããªãæ¹ã§ããï¼ããªãããããããªããã¨æã£ãããã°ãªã¼ã³ã¹ãã³ã¯ãçµæ¸å¦çã«ã¯ãããªããããå¾ãªãã ãããã¨è¨ã話ããã¦ããã®ã«ããããæ¿æ²»çãªè©±ã«æ´ç¨ãã¦ãã両è ããã£ã¡ãã«ãã¡ããã¡ã§ãã ã°ãªâ¦
ããæ²ããä¸å¤®ç·ãã沿ç·å¥ã»æ±äº¬è¿é101ã®è¡ãå± ä½å°ã¨ææéã®æãã¹ãé¢ä¿ããããã¹ã¼ãã¼ã¬ã¤ãããå°å³åãã¦ã¿ããé¢ç½ãåæãçµæ§ããªã¥ããåé¢ããã¯ãç°ç¡ã®åå¨ãæ°ã«ãªããåã¯è¡ã£ããã¨ãªããã ãã©ãã©ããªãã§ãããã Never Come Again â¦
IMFã®ãã´ãæ°ãå¾ã§è©³ããæ¸ããã©ãã¨ããããè¦ç¹ã ãã ã追è¨ãã¾ãããã¨ããããç®æ¡æ¸ãã§ãã ä¸çåæããã¬ã®ãªã¹ã¯ã¯ãåå¨ã¯ããã大ããã¯ãªã éèç·©åã®çµæã¨ãã¦ã®åå®ã¯å®¹èªãããã ãã æ¥éã5ã10å¹´ã®é·æã¤ã³ãã¬ç®æ¨ãæ²ãããã¨ã¯â¦