¸«Åö¤Î¤¿¤á¤ÎWeka¤ÎÁàºî¼ê½ç
£±¡¥Í×°ø(ÊÑ¿ô)¤ÎÁªÂò
£±¡Ë³µÍ×
ÌÜŪ¤Ï¡¢¤É¤ÎÍ×°ø¤òºîÍѤµ¤»¤ì¤ÐãÀ®¤Ç¤¤ë¤Î¤«¡©¤½¤ì¤ò¸«Äê¤á¤ë¤Î¤¬¡¢ÊÑ¿ô¤ÎÁªÂò¤Ç¤¹¡£
Ä̾ï¤Î¥Þ¡¼¥±¥Æ¥£¥ó¥°²ÝÂê¤Ç¤Ï¡¢£²¤Ä¤ÎÊýË¡¤ÇÌÜŪÊÑ¿ô¤ËÂФ¹¤ëºîÍѤòɾ²Á¤·¤Þ¤¹¡£
GeneticSearch¡¡¡¡¡¡¡¡ ¡§°äÅÁŪ¥¢¥ë¥´¥ê¥º¥à¤Ë¤è¤ëõº÷
GainRatioAttributeEval¡§¾ðÊóÍøÆÀÈæ¤Ë¤è¤ëɾ²Á
°äÅÁŪ¥¢¥ë¥´¥ê¥º¥à¤Ë¤è¤ëõº÷¤Ç¡¢ÌÜŪÊÑ¿ô¤ËÂФ·¤Æ¡Ö¸ú¤¤¤Æ¤¤¤ë¡×ÊÑ¿ô¤òʬ¤±¡¢¾ðÊóÍøÆÀÈæ¤Ç¡Ö¤É¤ÎÄøÅ١׸ú¤¤¤Æ¤¤¤ë¤«¤Çʤ٤ޤ¹¡£¡ØʬÀϤȤϡ¢Ê¬¤±¤Æʤ٤뤳¤È¡Ù¤Ç¤¹¡£
£²¡ËÁàºî
Select attributes¥¿¥Ö¤òÁªÂò
¢ª Choose¥Ü¥¿¥ó¤ò²¡²¼ ¢ª CfsSubsetEval¤òÁªÂò
¢ª Choose¥Ü¥¿¥ó¤ò²¡²¼ ¢ª GeneticSearch¤òÁªÂò
¢ª ¥×¥ë¥À¥¦¥ó¤ÇÌÜŪÊÑ¿ô¤òÁªÂò ¢ª Start¤ò²¡²¼
¢ª Choose¥Ü¥¿¥ó¤ò²¡²¼ ¢ª GainRatioAttributeEval¤òÁªÂò
¢ª Choose¥Ü¥¿¥ó¤ò²¡²¼ ¢ª Ranker¤òÁªÂò
¢ª ¥×¥ë¥À¥¦¥ó¤ÇÌÜŪÊÑ¿ô¤òÁªÂò ¢ª Start¤ò²¡²¼
£³¡Ë·ë²Ì
°äÅÁŪ¥¢¥ë¥´¥ê¥º¥à¤Ë¤è¤ëõº÷¤Ë¤è¤ê¡¢£µ¤Ä¤ÎÊÑ¿ô¤¬¡¢ÌÜŪ¤ËºîÍѤ¹¤ëÍ×°ø¤È¤·¤Æ¸õÊä¤Ëµó¤¬¤ê¤Þ¤·¤¿¡£
¾ðÊóÍøÆÀÈæ¤Ë¤è¤ëɾ²Á¤Ë¤è¤ê¡¢¸õÊä¤ËÁª¤é¤Ð¤ì¤¿£µ¤Ä¤ÎÍ×°ø¤¬¡¢ÁêÂÐŪ¤Ë¹â¤¤ÀâÌÀÎÏ(ÌÜŪ¤ËÂФ¹¤ëºîÍÑ)¤ò»ý¤Ã¤Æ¤¤¤ë¤³¤È¤¬³Îǧ¤µ¤ì¤Þ¤·¤¿¡£
Search Method: Genetic search¡Ê°äÅÁŪ¥¢¥ë¥´¥ê¥º¥à¤Ë¤è¤ëõº÷¡Ë
Probability of crossover: 0.6
¸òºµ³ÎΨ¤Ï¤É¤Î¤¯¤é¤¤¤ÎÉÑÅ٤Ǹòºµ¤¬¹Ô¤ï¤ì¤ë¤«¤ò¼¨¤·¤Æ¤¤¤Þ¤¹¡£¤â¤·¸òºµ¤¬µ¯¤³¤é¤Ê¤±¤ì¤Ð¡¢»Ò¹¤Ïξ¿Æ¤Î´°Á´¤Ê¥³¥Ô¡¼¤È¤Ê¤ê¤Þ¤¹¡£¤â¤·¸òºµ¤¬µ¯¤³¤ì¤Ð¡¢»Ò¹¤Ïξ¿Æ¼Á¤ÎÀ÷¿§ÂΤΰìÉôʬ¤Å¤Ä¤«¤é¤Ç¤¤¢¤¬¤ê¤Þ¤¹¡£¤â¤·¸òºµ³ÎΨ¤¬100%¤Ç¤¢¤ì¤Ð¡¢¤¹¤Ù¤Æ¤Î»Ò¹¤Ï¸òºµ¤Ë¤è¤ê¤Ç¤¤¢¤¬¤ê¤Þ¤¹¡£¤â¤·0%¤Ç¤¢¤ì¤Ð¡¢¿·¤·¤¤À¤Âå¤ÏÁ°¤Î¸ÄÂ粤ÎÀ÷¿§ÂΤδ°Á´¤Ê¥³¥Ô¡¼¤È¤Ê¤ê¤Þ¤¹¡£¡Ê¤·¤«¤·¤³¤ì¤Ï¿·¤·¤¤À¤Â夬Ʊ¤¸Êª¤Ç¤¢¤ë¤È¤¤¤¦¤³¤È¤Ç¤Ï¤¢¤ê¤Þ¤»¤ó!¡Ë
¸òºµ¤Ï¿·¤·¤¤À÷¿§ÂΤ¬¸Å¤¤À÷¿§ÂΤÎÎɤ¤Éôʬ¤ò¤â¤Ã¤Æ¤¤¤ë¤È¤¤¤¦´üÂԤȡ¢¿·¤·¤¤¤â¤Î¤¬¤è¤ê¤è¤¯¤Ê¤ë¤È¤¤¤¦´üÂÔ¤ò¹þ¤á¤Æ¤Ä¤¯¤é¤ì¤Þ¤¹¡£¤·¤«¤·¤Ê¤¬¤é¸ÄÂη²¤Î¤¤¤¯¤Ä¤«¤ÎÉôʬ¤ò¼¡¤ÎÀ¤Âå¤Ø»Ä¤¹¤Î¤â¤è¤¤¤³¤È¤Ç¤¹¡£
Probability of mutation: 0.033
ÆÍÁ³ÊÑ°Û³ÎΨ¤ÏÀ÷¿§ÂΤΰìÉô¤¬¤É¤ì¤¯¤é¤¤¤ÎÉÑÅÙ¤ÇÆÍÁ³ÊÑ°Û¤òµ¯¤³¤¹¤«¤ò¤¢¤é¤ï¤·¤Æ¤¤¤Þ¤¹¡£¤â¤·ÆÍÁ³ÊÑ°Û¤¬¤Ê¤±¤ì¤Ð»Ò¹¤Ï¤Ê¤ó¤ÎÊѹ¹¤â¤Ê¤·¤Ë¸òºµ¡Ê¤Þ¤¿¤Ï¥³¥Ô¡¼¡Ë¤òbe taken after ÆÍÁ³ÊÑ°Û¤¬¹Ô¤ï¤ì¤ë¤ÈÀ÷¿§ÂΤΰìÉôʬ¤¬Êѹ¹¤µ¤ì¤Þ¤¹¡£¤â¤·0%¤Ç¤¢¤ë¤ÈÊѹ¹¤µ¤ì¤Þ¤»¤ó¡£ÆÍÁ³ÊÑ°Û¤ÏGA¤¬¶É½ê²ò¤Ë´Ù¤ë¤Î¤òËɤ°¤¿¤á¤Ëºî¤é¤ì¤Æ¤¤¤Þ¤¹¡¢¤·¤«¤·¤¢¤Þ¤ê¤½¤ì¤ÏÉÑÈˤˤϤª¤³¤ê¤Þ¤»¤ó¡¢¤Ê¤¼¤Ê¤éGA¤Ï¼ÂºÝ¤Ï¥é¥ó¥À¥à¥µ¡¼¥Á¤ËÊѹ¹¤µ¤ì¤Æ¤¤¤ë¤«¤é¤Ç¤¹¡£
Attribute Subset Evaluator (supervised, Class (nominal): 17 FL):
Selected attributes: 2,3,5,10,15 : 5
F02
F03
F05
F10
F15
Search Method: Gain Ratio feature evaluator¡Ê¾ðÊóÍøÆÀÈæ¤Ë¤è¤ëɾ²Á¡Ë
Ranked attributes:
----------------
0.40679 15 F15
0.28652 10 F10
0.21274 3 F03
0.04417 2 F02
0.03473 5 F05
----------------
0.02176 9 F09
0.02093 1 F01
0.01803 6 F06
0.01625 13 F13
0.01609 7 F07
0.01374 12 F12
0.01329 14 F14
0.00815 8 F08
0.00413 11 F11
0.00122 4 F04
0 16 F16
£±¡¥Í×°ø(ÊÑ¿ô)¤ÎÁªÂò
£±¡Ë³µÍ×
ÌÜŪ¤Ï¡¢¤É¤ÎÍ×°ø¤òºîÍѤµ¤»¤ì¤ÐãÀ®¤Ç¤¤ë¤Î¤«¡©¤½¤ì¤ò¸«Äê¤á¤ë¤Î¤¬¡¢ÊÑ¿ô¤ÎÁªÂò¤Ç¤¹¡£
Ä̾ï¤Î¥Þ¡¼¥±¥Æ¥£¥ó¥°²ÝÂê¤Ç¤Ï¡¢£²¤Ä¤ÎÊýË¡¤ÇÌÜŪÊÑ¿ô¤ËÂФ¹¤ëºîÍѤòɾ²Á¤·¤Þ¤¹¡£
GeneticSearch¡¡¡¡¡¡¡¡ ¡§°äÅÁŪ¥¢¥ë¥´¥ê¥º¥à¤Ë¤è¤ëõº÷
GainRatioAttributeEval¡§¾ðÊóÍøÆÀÈæ¤Ë¤è¤ëɾ²Á
°äÅÁŪ¥¢¥ë¥´¥ê¥º¥à¤Ë¤è¤ëõº÷¤Ç¡¢ÌÜŪÊÑ¿ô¤ËÂФ·¤Æ¡Ö¸ú¤¤¤Æ¤¤¤ë¡×ÊÑ¿ô¤òʬ¤±¡¢¾ðÊóÍøÆÀÈæ¤Ç¡Ö¤É¤ÎÄøÅ١׸ú¤¤¤Æ¤¤¤ë¤«¤Çʤ٤ޤ¹¡£¡ØʬÀϤȤϡ¢Ê¬¤±¤Æʤ٤뤳¤È¡Ù¤Ç¤¹¡£
£²¡ËÁàºî
Select attributes¥¿¥Ö¤òÁªÂò
¢ª Choose¥Ü¥¿¥ó¤ò²¡²¼ ¢ª CfsSubsetEval¤òÁªÂò
¢ª Choose¥Ü¥¿¥ó¤ò²¡²¼ ¢ª GeneticSearch¤òÁªÂò
¢ª ¥×¥ë¥À¥¦¥ó¤ÇÌÜŪÊÑ¿ô¤òÁªÂò ¢ª Start¤ò²¡²¼
¢ª Choose¥Ü¥¿¥ó¤ò²¡²¼ ¢ª GainRatioAttributeEval¤òÁªÂò
¢ª Choose¥Ü¥¿¥ó¤ò²¡²¼ ¢ª Ranker¤òÁªÂò
¢ª ¥×¥ë¥À¥¦¥ó¤ÇÌÜŪÊÑ¿ô¤òÁªÂò ¢ª Start¤ò²¡²¼
£³¡Ë·ë²Ì
°äÅÁŪ¥¢¥ë¥´¥ê¥º¥à¤Ë¤è¤ëõº÷¤Ë¤è¤ê¡¢£µ¤Ä¤ÎÊÑ¿ô¤¬¡¢ÌÜŪ¤ËºîÍѤ¹¤ëÍ×°ø¤È¤·¤Æ¸õÊä¤Ëµó¤¬¤ê¤Þ¤·¤¿¡£
¾ðÊóÍøÆÀÈæ¤Ë¤è¤ëɾ²Á¤Ë¤è¤ê¡¢¸õÊä¤ËÁª¤é¤Ð¤ì¤¿£µ¤Ä¤ÎÍ×°ø¤¬¡¢ÁêÂÐŪ¤Ë¹â¤¤ÀâÌÀÎÏ(ÌÜŪ¤ËÂФ¹¤ëºîÍÑ)¤ò»ý¤Ã¤Æ¤¤¤ë¤³¤È¤¬³Îǧ¤µ¤ì¤Þ¤·¤¿¡£
Search Method: Genetic search¡Ê°äÅÁŪ¥¢¥ë¥´¥ê¥º¥à¤Ë¤è¤ëõº÷¡Ë
Probability of crossover: 0.6
¸òºµ³ÎΨ¤Ï¤É¤Î¤¯¤é¤¤¤ÎÉÑÅ٤Ǹòºµ¤¬¹Ô¤ï¤ì¤ë¤«¤ò¼¨¤·¤Æ¤¤¤Þ¤¹¡£¤â¤·¸òºµ¤¬µ¯¤³¤é¤Ê¤±¤ì¤Ð¡¢»Ò¹¤Ïξ¿Æ¤Î´°Á´¤Ê¥³¥Ô¡¼¤È¤Ê¤ê¤Þ¤¹¡£¤â¤·¸òºµ¤¬µ¯¤³¤ì¤Ð¡¢»Ò¹¤Ïξ¿Æ¼Á¤ÎÀ÷¿§ÂΤΰìÉôʬ¤Å¤Ä¤«¤é¤Ç¤¤¢¤¬¤ê¤Þ¤¹¡£¤â¤·¸òºµ³ÎΨ¤¬100%¤Ç¤¢¤ì¤Ð¡¢¤¹¤Ù¤Æ¤Î»Ò¹¤Ï¸òºµ¤Ë¤è¤ê¤Ç¤¤¢¤¬¤ê¤Þ¤¹¡£¤â¤·0%¤Ç¤¢¤ì¤Ð¡¢¿·¤·¤¤À¤Âå¤ÏÁ°¤Î¸ÄÂ粤ÎÀ÷¿§ÂΤδ°Á´¤Ê¥³¥Ô¡¼¤È¤Ê¤ê¤Þ¤¹¡£¡Ê¤·¤«¤·¤³¤ì¤Ï¿·¤·¤¤À¤Â夬Ʊ¤¸Êª¤Ç¤¢¤ë¤È¤¤¤¦¤³¤È¤Ç¤Ï¤¢¤ê¤Þ¤»¤ó!¡Ë
¸òºµ¤Ï¿·¤·¤¤À÷¿§ÂΤ¬¸Å¤¤À÷¿§ÂΤÎÎɤ¤Éôʬ¤ò¤â¤Ã¤Æ¤¤¤ë¤È¤¤¤¦´üÂԤȡ¢¿·¤·¤¤¤â¤Î¤¬¤è¤ê¤è¤¯¤Ê¤ë¤È¤¤¤¦´üÂÔ¤ò¹þ¤á¤Æ¤Ä¤¯¤é¤ì¤Þ¤¹¡£¤·¤«¤·¤Ê¤¬¤é¸ÄÂη²¤Î¤¤¤¯¤Ä¤«¤ÎÉôʬ¤ò¼¡¤ÎÀ¤Âå¤Ø»Ä¤¹¤Î¤â¤è¤¤¤³¤È¤Ç¤¹¡£
Probability of mutation: 0.033
ÆÍÁ³ÊÑ°Û³ÎΨ¤ÏÀ÷¿§ÂΤΰìÉô¤¬¤É¤ì¤¯¤é¤¤¤ÎÉÑÅÙ¤ÇÆÍÁ³ÊÑ°Û¤òµ¯¤³¤¹¤«¤ò¤¢¤é¤ï¤·¤Æ¤¤¤Þ¤¹¡£¤â¤·ÆÍÁ³ÊÑ°Û¤¬¤Ê¤±¤ì¤Ð»Ò¹¤Ï¤Ê¤ó¤ÎÊѹ¹¤â¤Ê¤·¤Ë¸òºµ¡Ê¤Þ¤¿¤Ï¥³¥Ô¡¼¡Ë¤òbe taken after ÆÍÁ³ÊÑ°Û¤¬¹Ô¤ï¤ì¤ë¤ÈÀ÷¿§ÂΤΰìÉôʬ¤¬Êѹ¹¤µ¤ì¤Þ¤¹¡£¤â¤·0%¤Ç¤¢¤ë¤ÈÊѹ¹¤µ¤ì¤Þ¤»¤ó¡£ÆÍÁ³ÊÑ°Û¤ÏGA¤¬¶É½ê²ò¤Ë´Ù¤ë¤Î¤òËɤ°¤¿¤á¤Ëºî¤é¤ì¤Æ¤¤¤Þ¤¹¡¢¤·¤«¤·¤¢¤Þ¤ê¤½¤ì¤ÏÉÑÈˤˤϤª¤³¤ê¤Þ¤»¤ó¡¢¤Ê¤¼¤Ê¤éGA¤Ï¼ÂºÝ¤Ï¥é¥ó¥À¥à¥µ¡¼¥Á¤ËÊѹ¹¤µ¤ì¤Æ¤¤¤ë¤«¤é¤Ç¤¹¡£
Attribute Subset Evaluator (supervised, Class (nominal): 17 FL):
Selected attributes: 2,3,5,10,15 : 5
F02
F03
F05
F10
F15
Search Method: Gain Ratio feature evaluator¡Ê¾ðÊóÍøÆÀÈæ¤Ë¤è¤ëɾ²Á¡Ë
Ranked attributes:
----------------
0.40679 15 F15
0.28652 10 F10
0.21274 3 F03
0.04417 2 F02
0.03473 5 F05
----------------
0.02176 9 F09
0.02093 1 F01
0.01803 6 F06
0.01625 13 F13
0.01609 7 F07
0.01374 12 F12
0.01329 14 F14
0.00815 8 F08
0.00413 11 F11
0.00122 4 F04
0 16 F16
·èÄê·×»»£± ¡Ý ¤½¤Î¿ô³Ø¤¬Àïά¤ò·è¤á¤ë
·èÄê·×»»£² ¡Ý ¤½¤Î¿ô³Ø¤¬Àïά¤ò·è¤á¤ë
·èÄê·×»»£³ ¡Ý ¤½¤Î¿ô³Ø¤¬Àïά¤ò·è¤á¤ë
·èÄê·×»»£´ ¡Ý ¤½¤Î¿ô³Ø¤¬Àïά¤ò·è¤á¤ë
·èÄê·×»»£µ ¡Ý Áê´Ø²òÀÏ
·èÄê·×»»£¶ ¡Ý £×£Å£Ë£Á¤Î³µÍ×
·èÄê·×»»£· ¡Ý Weka¤ò»È¤ª¤¦
·èÄê·×»»£¸ ¡Ý WEKA¤ò»È¤¦£²
·èÄê·×»»£¹ ¡Ý Weka¤ò»È¤¦£³
·èÄê·×»»£±£° ¡Ý Weka¤ò»È¤¦£´
·èÄê·×»»£±£± ¡Ý Weka¤ò»È¤¦£µ
·èÄê·×»»£±£² ¡Ý ¥Ç¡¼¥¿¥Þ¥¤¥Ë¥ó¥°ºÆ¹Í
·èÄê·×»»£±£³ ¡Ý ¿½Å¶¦ÀþÀ
·èÄê·×»»£±£´ ¡Ý WEKA¤ÎÁàºî¼ê½ç£±
·èÄê·×»»£±£µ ¡Ý WEKA¤ÎÁàºî¼ê½ç£²
·èÄê·×»»£±£¶ ¡Ý WEKA¤ÎÁàºî¼ê½ç£³
·èÄê·×»»£±£· ¡Ý WEKA¤ÎÁàºî¼ê½ç£´
·èÄê·×»»£±£¸ ¡Ý °ìÈÌÀþ·Á¥â¥Ç¥ë