2019-10-01ãã1ã¶æéã®è¨äºä¸è¦§
ã¨ããè¨äºãè¦ã¦ãèªåã®ããã°ãå·çæ°ã¯ããããããã®ã§è¯ãç·è¡ãã®ã§ã¯ãªããã¨æã£ã¦éè¨ãã¦ã¿ã¾ãããåã®å人ããã°ããããããè¯ãã¨ããè¡ãããããªããããã===150社ã®Techããã°ãåæãã¦è¦ãããã¨ã³ã¸ãã¢ãä»è»¢è·ããã¹ãä¼æ¥ã©ã³ãã³â¦
Kaggleãªã©ã®ãã¼ã¿åæã³ã³ãã§ãã¼ãã«ãã¼ã¿ãæ±ãå ´åãæè¿ã¯åããããLightGBMãå©ç¨ããå ´åãå¤ãã§ããæ¬è¨äºã§ã¯ãåæã®æ©æ¢°å¦ç¿ã¢ã«ã´ãªãºã ã¨ãã¦ãLightGBMã*1ãæ¡ç¨ããçç±ãç´¹ä»ãã¾ããããã¾ã§2019å¹´10ææ«æç¹ã§ã®å人ã®ä¸»è¦³ãªã®ã§ãâ¦
å æ¥å ¬éãããIEEE-CIS Fraud Detectionãã³ã³ãã®è§£æ³*1ã®ä¸ã§ãAdversarial Validationã®èãæ¹ãç¨ããç¹å¾´éé¸æã«ã¤ãã¦ä½åã質åãããã¾ãããæ¬è¨äºã§ã¯ãAdversarial Validationã®èãæ¹ãç¨ããç¹å¾´éé¸æã解説ãã¾ãã Adversarial Validationâ¦
ãã¼ã¿ãµã¤ã¨ã³ãã£ã¹ãåä¼ 6thã·ã³ãã¸ã¦ã ãKaggle Grandmasterã«èãï¼ï¼ããããã¼ã¿ãµã¤ã¨ã³ãã£ã¹ãã®éå»ã»ç¾å¨ã»æªæ¥ï¼ãã«ãã¢ãã¬ã¼ã¿ã¼ã¨ãã¦ç»å£ãã¾ãããwww.datascientist.or.jpç§ãGrandmasterã®JackãããOnoderaããã«è³ªåããå½¢å¼ã®ããâ¦
KaggleãIEEE-CIS Fraud Detectionãã³ã³ãã«å人ã§åå ãã¦ã2485ä½ã§ãããpublic lb ã¹ã³ã¢ã ã¨2800ä½ç¨åº¦ã®æåºã§shake upãçã£ãã®ã§ãã妥å½ãªçµæã«çµãã£ã次第ã§ããæ¬è¨äºã§ã¯ãdiscussionã«æ稿ããå 容ãåºã«ãæ¬ã³ã³ãã§ã®åãçµã¿ãã¾ã¨ãã¾â¦
éå®ã®å è¡è²©å£²*1ã§ç´çãå ¥æãããKaggleã§åã¤ãã¼ã¿åæã®æè¡ãï¼æè¡è©è«ç¤¾ï¼ãèªã¿ã¾ããããªãé»åçããæµè´ããã ãäºå®ã§ããgihyo.jp10æ7æ¥ã®çºå£²ãå¾ ããã㦠Amazon*2ã®ãã¹ãã»ã©ã¼1ä½ã«ãªããªã©ã注ç®ãéãã¦ãã¾ããæ¢ã«èè ã®ä¸äººã§ããâ¦