åã«æ¸ããSVMã®è¨äºã§ããL1ã¨ãL2ã¨ããã®ã¯ééããã¨ãã®ããã«ãã£ãã©ãå®ç¾©ããããæå³ãã¦ãããã¨æ¸ãã¦ããããL1ã¨ãL2ã£ã¦æ£ååé ã®è©±ãªããããªãã®ãã¨çåã«æã£ãã1ã¶æã»ã©æéãããã¦ã®ã»ã«ãããã³ãã§ããã確èªãããã¨ãã¦ã«ã¼ãã«å¤å¤é解æãèªãã¨ããã¯ãæ£ååé ã«ã¤ãã¦ã¯L1ã¨L2ã®ä¸¡æ¹ã®èª¬æãæ¸ãã¦ããããæ失ã«é¢ãã¦ã¯æ®éã®Hinge Lossï¼=L1 Lossï¼ããæ¸ãã¦ãªãã ã¨è¨ã訳ã§ããããééãã¡ãã£ããªããã¨ææ¾¹ããæ°æã¡ã§"A dual coordinate descent method for large-scale linear SVM"ãèªã¿ç´ãã¦ã¿ãã¨ããããã£ã±ãL1-SVMã¨ããã®ã¯æ失ãæ®éã®Hinge Lossã§ãL2-SVMã¨ããã®ã¯Hinge Lossã®2ä¹ãæ失ã¨ããã¨æ¸ãã¦ãã£ãã両æ¹ã¨ãæ£ååé ã«ã¤ãã¦ã¯L2æ£ååã使ã£ã¦
{{#tags}}- {{label}}
{{/tags}}