full2010.pdf - page 1517

1479
‡ÎµœÎµ
„µ¦®µ‡Îµ˜°ÄœÁ¦ºÉ°ŠÄ—Á¦ºÉ°Š®œ¹ÉŠ °µ‹„¦³šÎµÅ—o ×¥„µ¦šÎµ„µ¦š—¨°ŠÁ¡ºÉ °´ŠÁ„˜Ÿ¨š¸Éŗo „µ¦š—¨°Šš¸É
šÎµŽÊε„´œ®¨µ¥Ç‡¦´ÊŠ£µ¥Ä˜o £µª„µ¦–rÁ—¸¥ª„´œ ×¥Ÿ¨š¸ÉÁž}œÅ—oš´ÊŠ®¤— °Š„µ¦š—¨°ŠÁ®¤º°œ„´œš»„‡¦´ÊŠ Ž¹ÉŠŸ¨ °Š
„µ¦šÎµŽÊε˜n¨³‡¦´ÊŠ°µ‹Áž}œŸ¨¨´¡›rėŸ¨¨´ ¡›r®œ¹ÉŠ °ŠŸ¨¨´¡›rš¸É°µ‹Áž}œÅžÅ—o Ĝ„µ¦š—¨°Šœ´ÊœŸ¨š¸Éŗo‹³Â˜„˜nµŠ
„´œÅ¤nÁ®¤º°œ„´œš»„‡¦´ÊŠ š´Ê Šœ¸ÊŸ¼oš—¨°ŠÅ¤n µ¤µ¦™¡¥µ„¦–r Ÿ¨¨´¡›rš¸ÉÁ„·— ¹ÊœÅ—o°¥nµŠÂœnœ°œ„n°œš¸É ‹³šÎµ„µ¦š—¨°Š
Á¦¸¥„„µ¦š—¨°Š¨´„¬–³Án œœ¸Êªnµ „µ¦š—¨°ŠÁ· Š»n¤ (Random Experiment) (µ¥¨ · œ¤¼¦–r š°Š, 2546)
„µ¦Â‹„‹ŠÁ¦ µ‡–·˜ Áž}œ„µ¦š—¨°ŠÁ·Š»n¤š¸É¤¸„µ¦š—¨°ŠŽÊε„´œÅžÁ¦ºÉ°¥ÇĜ¦¼ž °Š‹Îµœªœ‡¦´ÊŠ °Š
‡ªµ¤ÎµÁ¦È‹®¦º °Å—o ·ÉŠš¸É˜o°Š„µ¦ (Success:p) ‹œ„ªnµ‹³ž¦³‡ªµ¤Å¤n εÁ¦È‹®¦º°Å—o ·ÉŠš¸ÉŤn˜o°Š„µ¦ (Failure) Áž}œ
‡¦´ÊŠÂ¦„‹¹Š®¥»—šÎµ„µ¦š—¨°Š ×¥„µ¦š—¨°Š»n¤Â˜n¨³‡¦´ÊŠÁž}œ°· ¦³‹µ„„´œ x
x f x p q; x 0,1, 2,...
˜´ªž¦³¤µ–宦´ ¡µ¦µ¤·Á˜°¦r°Š„µ¦Â‹„‹ŠÁ¦ µ‡–·˜ ‡º° ˆp x 1 x —´Šœ´Êœ ˜´ªž¦³¤µ–宦´‡nµÁŒ¨¸É¥
°Š„µ¦Â‹„‹ŠÁ¦ µ‡–·˜ ‡º° ˆ ˆ ˆ p q x P
„µ¦°œ»¤µœ ®¦º°¦»ž‡»–¨´„¬–³ °Šž¦³µ„¦ÂÁ¦ µ‡–·˜ µ¤µ¦™šÎ µÅ—o °Š¨´ „¬–³‡º° „µ¦
ž¦³¤µ–‡nµ (Estimation) ¨³„µ¦š—°¤¤˜·“µœ (Test of hypothesis) Ĝ„µ¦ª·‹´¥‡¦´ÊŠœ¸Ê œÄ‹«¹„¬µÁ„¸É¥ª„´„µ¦
ž¦³¤µ–宦´ ‡nµÁŒ¨¸É¥ °Š„µ¦Â‹„‹ŠÁ¦ µ‡–·˜ Ž¹ÉŠšÎµÅ—o 2 ¨´„¬–³‡º° „µ¦ž¦³¤µ–‡nµÂ‹»— (Point
Estimation) ¨³„µ¦ž¦³¤µ–‡nµÂnªŠ (Interval Estimation) Ž¹ÉŠÁ¦µ‹³¡·‹µ¦–µÁŒ¡µ³„µ¦ž¦³¤µ–‡n µÂnªŠ
Ášnµœ´Ê œ ץĜ„µ¦ž¦³¤µ–‡nµÂnªŠÎµ®¦´‡nµÁŒ¨¸É¥ P ¦¼žÂ®œ¹É Šš¸Éœ· ¥¤Äo„´œš´ÉªÅžÄœ„µ¦ž¦³¤µ–‡nµ
nªŠ 宦´ ‡nµÁŒ¨¸É¥ °Š„µ¦Â‹„‹ŠÁ¦ µ‡–·˜ Á¤ºÉ°„ε®œ—´¤ž¦³·š›·Í‡ªµ¤ÁºÉ°¤´É œ (Confidence coefficient)
1 D ༡
2
1 2
ˆ ˆ
ˆ Z
n D
P P
P P r
Á¤ºÉ° 1 2 Z D ‡º° ‡ª°œÅš¨rš¸É 1 2D °Š„µ¦Â‹„‹Šž„˜·¤µ˜¦“µœ ¨³ ˆ x P
„µ¦ž¦³¤µ–‡nµÂnªŠÄœ¦¼žÂ oµŠ˜oœœ¸ÊÁ¦¸¥„ªnµ ª·›¸ž„˜· ˜n¥´ŠÄ®o„µ¦‡¦°‡¨»¤‡nµÁŒ¨¸É¥š¸ÉŤn—¸œ´„
Šµœª·‹´¥œ¸Ê ‹¹ŠœÄ‹ª·›¸„µ¦ž¦³¤µ–‡nµÂnªŠÎµ®¦´‡nµÁŒ¨¸É¥ °Š„µ¦Â‹„‹ŠÁ¦ µ‡–·˜š¸ÉÄ®o„µ¦‡¦°‡¨»¤‡nµÁŒ¨¸É¥š¸É
Á®¤µ³¤„ªnµª·›¸ž„˜· ‹µ„„µ¦˜¦ª‹°Á°„µ¦¡ªnµ ¥´ Š¤¸ª·›¸š¸Éœn µœÄ‹°ºÉœ°¸„Ánœ
Brown et al. (2003). «¹„¬µª·›¸„µ¦¦oµŠnªŠ‡ªµ¤ÁºÉ°¤´ÉœÎµ®¦´‡nµÁŒ¨¸É¥ °Š„µ¦Â‹„‹ŠÁ¦ µ‡–·˜—oª¥
ª·›¸ ‡°¦r ¨³ª· ›¸°´˜¦µnªœ£µª³œnµ‹³Áž}œ ×¥Áž¦¸¥Áš¸¥„´ª· ›¸ž„˜· ¡ªnµ ª·›¸ ‡°¦r¨³ª·›¸°´˜¦µnªœ£µª³œnµ‹³
Áž}œÁž}œª· ›¸š¸É¤¸‡ªµ¤Á®¤µ³¤„ªnµª·›¸ž„˜· Á¤ºÉ°¡·‹µ¦–µ™¹Š‡ªµ¤œnµ‹³Áž}œš¸ÉnªŠ‡ªµ¤ÁºÉ°¤´É œ‹³‡¦°‡¨»¤‡nµÁŒ¨¸É¥
¨³Äoŗo „´š»„ œµ—˜´ª°¥nµŠš¸ÉšÎµ„µ¦«¹„¬µ
ÁœºÉ°Š‹µ„„µ¦ž¦³¤µ–‡nµÃ—¥ª·›¸ ž„˜·Â¨³ª·›¸ ‡°¦r čo˜´ªž¦³¤µ–‡nµÂ‹»—宦´‡nµÁŒ¨¸É¥ ˆ x P
—oª¥ª·›¸£µª³œn µ‹³Áž}œ¼Š»— ‹¹ŠÅ—o¤¸„µ¦Áœ°Âœª‡·—Ä®¤nץčo˜´ªž¦³¤µ–Ár (Bayes Estimator) ×¥
„ε®œ—¢{Š„r´œ‡ªµ¤®œµÂœnœÁ¦·É¤Â¦„ (Prior density function) Áž}œ„µ¦Â‹„‹ŠÁ°„¦¼ž (Uniform Distribution)
ĜnªŠ (0,1) —´Šœ´Êœ‹³Å—o˜´ªž¦³¤µ–‡nµÁŒ¨¸É¥˜´ªÄ®¤n‡º° x 1
n 1
P ¦ ¨³Äo ‡nµÁŒ¨¸É¥ P šœ‡nµÁŒ¨¸É¥š¸Éž¦³¤µ–
ŗo‹µ„ª· ›¸£µª³œnµ‹³Áž}œ¼Š»—Äœª·›¸ž„˜· ¨³ª·›¸ ‡°¦r ×¥Á¦¸¥„ªnµ ª·›¸ ž„˜·ÂÁr (Bayes Normal method) ¨³
ª·›¸ ‡°¦rÁr (Bayes Score method) ˜µ¤¨Îµ—´