Çë½ÌROC curveÓë gians chart

ÊÇ ·ñ +2 ÂÛ̳±Ò

kÈË ²ÎÓë»Ø´ð

¾­¹ÜÖ®¼ÒËÍÄúÒ»·Ý

Ó¦½ì±ÏÒµÉúרÊô¸£Àû!

ÇóÖ°¾ÍҵȺ

ÕÔ°²¶¹ÀÏʦ΢ÐÅ£ºzhaoandou666

¾­¹ÜÖ®¼ÒÁªºÏCDA

ËÍÄúÒ»¸öÈ«¶î½±Ñ§½ðÃû¶î~ !

Á¢¼´ÁìÈ¡

¸ÐлÄú²ÎÓëÂÛ̳ÎÊÌâ»Ø´ð

¾­¹ÜÖ®¼ÒËÍÄúÁ½¸öÂÛ̳±Ò£¡

+2 ÂÛ̳±Ò

Á½Õ߶¼¿ÉÒÔÓÃÀ´validate logistic model×öÐÔÄÜÆÀ¹À£¬²»Ì«Ã÷°×Á½ÕߵĹØϵ¡£ÊDz»ÊÇ¿ÉÒÔ´Óroc curve ÖеÄsensitivity and specifity °Ñresponse rate and lift value Ëã³öÀ´¡£

ɨÂë¼ÓÎÒ À­ÄãÈëȺ

Çë×¢Ã÷£ºÐÕÃû-¹«Ë¾-ְλ

ÒÔ±ãÉóºË½øȺ×ʸñ£¬Î´×¢Ã÷Ôò¾Ü¾ø

·ÖÏí0 ÊÕ²Ø0 »ØÌû

¹Ø¼ü´Ê£ºcurve gians Chart Hart ans ÆÀ¹À

We prove a relation between the accuracy ratio and the area (A) under the ROC curve in order to demonstrate that both measures are equivalent. By a simple calculation, we get the connection:
GINI=2(A-0.5)
Note: A is area under ROC curve.