NLP
Precision , Recall, F-measure, R-Precision çã®çµæ: t(true) or f(false) æ¤ç´¢çµæ: p(positive) or n(negative) åç¾ç(Recall) é©åç(Precision) Få¤(F-measure) â»Få¤ã¯åç¾çã¨é©åçã®èª¿åå¹³å ãã¤ã¯ãå¹³å(micro average)ï¼ã·ã¹ãã æåã®å°ºåº¦(ã¯â¦
ã¨ã³ãããã¼(entropy) ã¨ã³ãããã¼(entropy)ï¼ç¢ºçå¤æ°ãã©ã®å¤ãåãããè¨ãå½ã¦ã«ãã(ä¹±éã)ãä¸ç¢ºããã測ã尺度ã 確çå¤æ°ãæã¤é¢æ£åå¸ã«å¯¾ãã¦ã確çå¤æ°ã®ã¨ã³ãããã¼ã¯ â»ã¯ããã¿ã«ã®å®çãã? ããã§ãã®ãã¹ã®ä¸çå¼ãããã¨ããã°ã ã¨ã³â¦
åºæ¬çã«é£ç¶ã§nåå¾®åå¯è½ãä»®å®ãã¦ããã ç®çé¢æ°(objective function) s.t. ä¸çå¼å¶ç´(inequality constraint) çå¼å¶ç´(equality constraint) å®è¡å¯è½è§£(feasible solution) å¶ç´ãæºãã解 å®è¡å¯è½é å(feasible region) å®è¡å¯è½è§£ã®éå 解ãå â¦