궷쁉뫞뉺뻇돸 닄꒭ꣷ 닄ꑔ듁 2003꙾11ꓫ Journal of Risk Management Nov. 2003 -361 맘쁉ꭏ돦ꚭ듁ꖢ껄꒧륷듺 Predicting Early Lapse of Life Insurance Policies * **뎯ꯘ돓 (Jian-Shen Chen) ꩌꧺꞻ (Ming-Horng Lin) 멋₭? ꖻ곣ꡳꝑꗎ룪껆뇄쑱덎꒤ꪺ엞뿨꽓끪쉫뭐쏾꾫롧뫴룴꣓륷듺맘쁉ꭏ돦ꚭ듁ꖢ껄꒧뻷뉶ꅃ뇄ꗎBoritz and Kennedy (1995)ꪺ뭾Ꝑꚨꖻ꣓뿅뙱륷듺볒ꮬꪺ쁵Ꙉ꧊ꅁ곣ꡳ떲ꩇ땯뉻ꅁ륂ꗎ귋뛇뮼쏾꾫롧뫴룴ꕩꕈꚳ껄ꙡ룑ꡍ엞뿨꽓끪쉫ꕈ엧ꭥ뻷뉶낵결셻곉귈ꛓ늣ꗍꪺꮬI뿹뭾륌ꑪꪺ냝썄ꅁꣃ궰ꝃ뭾Ꝑꪺ띬ꖢꚨꖻꅃ꣤ꚸꅁꝑꗎ엞뿨꽓끪쉫꧒뽺뿯ꪺ엣뗛엜볆ꅁꕩꕈ룑ꡍ귋뛇뮼쏾꾫롧뫴룴때ꩫ뮡ꧺ엜볆뚡ꪺꙝꩇ쏶ꭙ꒧냝썄ꅃ낣ꚹ꒧ꕾꅁ귋뛇뮼쏾꾫롧뫴룴꒧뭾Ꝑꚨꖻꣃ꒣라쁈뗛ꚨꖻ뉶ꪺ뱗ꕛꛓꚳꓓꑪꪺ엜꓆ꅃ돌ꯡꅁ롧ꗑ맪뗽녯ꪾ띾냈ꑈ귻걏ꝟ룵병싷슾ꅂꩁ냈ꭾ뷨ꅂ쎺뙏ꓨꚡ뭐ꭏ뙏굴뻡꿠ꑏ떥ꙝ꿀맯ꭏ돦걏ꝟꚭ듁ꖢ껄ꚳꯜ엣뗛ꪺ뱶암ꅃ 쏶쇤꙲ꅇ쏾꾫롧뫴룴ꅂ엞뿨꽓끪쉫ꅂ셻곉귈ꅂꮬI뿹뭾ꅂ뭾Ꝑꚨꖻꅃ Abstract This paper applies data mining technology to predict the probability of early lapse for life insurance policy. Based on the misclassification cost of Boritz and Kennedy (1995), we compared the performance between the neural network and logistic regression models. The results of the study indicate that, firstly, back-propagation neural network can solve the problem of logistic regression that use prior probability for cut-off value that result a higher type I error. Secondly, by using the significant explanatory variables obtained from the result of logistic regression, it is helpful to explain the relationship between output and input variables of back-propagation neural network model. Moreover, misclassification cost of the back-propagation neural network is more stable than others in varied cost ratios. Finally, the most important factors that affect the probability of early lapse are the problem of salesmen’s quitting job, the service quality of business affairs personnel, premium payment mode and ability to pay premium. KeywordsꅇNeural Network, Logistic Regression, Cut-off Value, Type I Error, Misclassification Cost. 띐쇂꣢ꛬ냎ꙗ뱦걤ꥥ귻뒣꣑ꪺ쑟뙑띎ꢣ꣣엩귗ꖿꯘ쒳ꅃ * 듂뚧곬ꑪ뻇ꭏ쁉뿄뫞뉺꡴뇐뇂ꅁProfessor, Department of Insurance, Chaoyang University of Technology, Taichung, Taiwan. ** 듂뚧곬ꑪ뻇ꭏ쁉뿄뫞뉺곣ꡳ꧒곣ꡳꗍꅁGraduate Student, Department of Insurance, Chaoyang University of Technology, Taichung, Taiwan.?
342맘쁉ꭏ돦ꚭ듁ꖢ껄꒧륷듺 1. ꭥꢥ 맘쁉ꭏ돦ꚭ듁ꪺꖢ껄뉶ꅝearly lapse rateꅞ걏뗻믹맘쁉띾냈ꭾ뷨ꪺ궫굮볐ꅁ꣤맯맘쁉꒽ꕱꪺ뱶암곛럭뉠뮷ꅃ냪맘쁉ꖫ돵Ꙣ1993꙾ꗾ궱뙽꧱띳맘쁉꒽ꕱꪺ덝ꗟꯡꅁꙝ결띳꒽ꕱ맯띾냈ꑈꑾ뭐ꪾꙗꯗꪺ믝ꡄꛓꓞ썺ꪺ꯵ꢤ궷볉ꅁꑝ덹ꚨ덜Ꙩ덳ꑈ녡ꗳꪺ룵병ꗳ꒣쉟ꙡꑗ면ꅁ뙩ꛓꭐ꣏냪맘쁉ꭏ돦ꪺꚭ듁ꖢ껄뉶꯹쓲ꙡ둣꓆ꅃ결떽ꚹꑀ냝썄ꅁ끝걆뎡ꭏ쁉ꕱ꧳1996꙾맪걉ꅵ띾냈롧샧ꛛꯟ럇ꭨ멛엳엩꣮떽군릺ꅶꅁ굮ꡄ냪꒺맘쁉꒽ꕱ꓀뚥걱륆ꚨꩫꥷꪺ닄ꑑꑔ귓ꓫ뭐닄ꑇꑑ꒭귓ꓫꭏ돦쑾쓲뉶ꓴ럇ꅁ뒫ꕹ룜뮡ꅁꭏ돦ꚭ듁ꪺꖢ껄뉶ꅝ1-쑾쓲뉶ꅞꖲ뚷놱꣮ꙢꙘ뉺ꪺ뵤돲꒺ꅁꝟꭨ뫊뉺뻷쏶녎꒣라껖귣룓꒽ꕱ띳ꭏ쁉냓ꭾꪺ땯ꛦꅃ맪걉ꛜ1999꙾꦳ꓮꪺ닎군볆뻚엣ꗜꅁꕈꗳ볆군뫢꒧맘쁉ꭏ돦닄ꑑꑔ귓ꓫ뭐닄ꑇꑑ꒭귓ꓫꪺꖢ껄뉶ꅁꑷ녱1996꙾ꪺ26%뭐33%ꑕ궰ꛜ20%뭐31%ꅁ맘쁉띾교ꑏ꧳떽맘쁉ꭏ돦ꚭ듁ꖢ껄뉶ꑷꚳꫬꡂꪺꚨ껄ꅃ쁈뗛냪ꕛꑊꕀ곉뙔꧶닕슴뭐룪ꮬꭏ돦ꪺ빐냢ꅁꙢꖫ돵쑶ꪧ뭐ꕾ냓ꭏ쁉꒽ꕱꪺꑪ셼ꑊꭉꑕꅁꖼ꣓걏ꝟ라ꙁꚸ덹ꚨꕴꑀ룑곹ꪺ궷볩ꅁ귈녯띾뭐뫊뉺뻷쏶ꪺ궫뗸ꅃ 껚뻚맘쁉꒽라닎군룪껆땯뉻ꅁꭏ돦ꗍ껄ꯡ꣢꙾꒺ꪺꖢ껄뉶돌낪ꅁ륌낪ꪺꭏ돦ꚭ듁ꖢ껄뉶맯꧳맘쁉꒽ꕱ뷄삻돌ꑪꅁ럭땍ꑝ걏ꭏ쁉뫊뉺뻷쏶ꪺ쏶ꩠ땊쉉ꅃꕄ굮ꪺ귬ꙝꙢ꧳꙾듁맘쁉ꭏ돦ꪺ띳ꮴ곹ꚨꖻꑪ뎣뚰꒤Ꙣ꧓ꭏꫬ듁꧒교ꅆꕝ걁띾냈ꑈ귻ꪺꅂ땯돦뙏ꗎꅂ엩샋뙏ꗎ빐냢뫞뉺뙏ꗎ떥ꅁꙝꚹ꙾듁맘쁉ꭏ돦Ꙣꖼ륆띬꽱ꖭ뿅쉉ꭥꝙꖢ껄ꅁ녎꣏녯ꭏ쁉꒽ꕱ때ꩫ앵Ꙟ띳ꮴ곹ꚨꖻꛓ덹ꚨ띬ꖢꅁ뙩ꛓ듮ꓖ맘쁉꒽ꕱꪺꝑ랽ꅃ맯덑ꭏ쁉ꑈꛓꢥꅁ낣돠ꖢ귬쎺ꭏ뙏꒧롧샙띬ꖢꕾꅁꗧ녎돠ꖢ귬ꚳꪺ맘쁉ꭏ믙ꅃꙝꚹꅁꭏ돦ꚭ듁ꖢ껄꒣덹ꚨꭏ돦꯹ꚳꑈ뭐ꭏ쁉꒽ꕱ싹ꓨꪺ롧샙띬ꖢꅁ쁈꒧돠ꖢꪺꭏ쁉ꭏ믙ꟳꕛ궫ꫀ라뫖ꝑꚨꖻꅃꙝꚹ꙰꛳ꚭ땯뉻ꝙ녎ꖢ껄ꪺꭏ돦ꅁ쇗ꝋ귬ꚳꭏꓡꙝ띾냈ꑈ귻ꩁ냈ꭾ뷨꒣꣎꒣럭ꪺ룜덎꒧뱶암ꛓ뮴꧶룑곹ꅁ녎걏맘쁉꒽ꕱꚳ껄ꙡ뫻꯹낪ꭏ돦쑾쓲뉶궰ꝃꭏ돦ꖢ껄뉶ꪺ뽮랥ꓨꩫꅁ덯ꑝ걏ꖻ곣ꡳꪺꕄ굮ꗘꪺꅃ 룪껆뇄쑱ꅝdata miningꅞꪺ띎롱걏결ꑆ땯뉻ꚳ띎롱ꪺ볒ꚡ덗ꭨꅁꕈꛛ냊ꕢꛛ냊ꪺꓨꚡ꣓냉걤ꅂ꓀꩒ꑪ뙱룪껆꧒뙩ꛦꪺ걹땻ꅝ둞ꓥꖿꅁ2001ꅞꅃ룪껆뇄쑱ꪺ믹귈ꝙꙢ꧳꙰꛳녱ꙕ꒽ꕱ꧒ꯘꗟꪺ썥ꑪ압ꯈ룪껆깷꒤꯵놸ꕘꚳ믹귈ꪺꡍ떦룪끔ꅃ맘쁉롧샧ꪺ냲ꖻ귬뉺결ꑪ볆ꩫꭨꅝlaw of large numberꅞ꒧륂ꗎꅁ꣤꽓쉉Ꙣ꧳꙰꛳ꑪ뙱ꙡ뫻꯹듁ꛓ쎭ꥷꪺꚳ껄ꮴ곹ꅁ꣏녯ꙝꚬ꣺ꪺꭏ뙏꧒뭅뚰꒧룪꿠냷ꕒ꓀륂ꗎ꧳룪ꅁꕈ샲녯돌ꑪꪺꝑ볭ꅃꙝꚹ꙰꛳녱ꑪ뙱ꪺꭏ돦룪껆깷꒤쉞꣺ꕘꚭ듁ꖢ껄ꕩ꿠꧊룻낪ꪺ맘쁉ꭏ돦ꅁ녎ꚳꝕ꧳띾냈ꭾ뷨뭐ꭏ돦쑾쓲뉶꒧뒣ꅃ쏶꧳꓀쏾냝썄ꪺ곣ꡳꅁ엞뿨
궷쁉뫞뉺뻇돸 343닄꒭ꣷ 닄ꑔ듁 2003꙾11ꓫ 꽓끪쉫ꅝlogistic regressionꅁ슲뫙LRꅞ뭐쏾꾫롧뫴룴ꅝneural networkꅞ곒결뒶륍삳ꗎ꧳럇ꭨ엜볆결ꑇ꓀쏾ꪺ곣ꡳꓨꩫꅁ꣢ꪺ깴늧Ꙣ꧳ꭥꕈ뵵꧊ꪺꓨꚡꕛꕈ꓀쏾ꅁꯡꕈꭄ뵵꧊ꪺꓨꚡ꣓꓀쏾ꅁ놩륂ꗎ쏾꾫롧뫴룴꧳맘쁉ꭏ돦ꚭ듁ꖢ껄꒧륷듺ꪺꓥ쑭ꣃ꒣Ꙩꢣꅃ냲꧳ꕈꑗ뉺ꗑꅁ결꿠ꚭꙝ삳ꭏ돦ꚭ듁ꖢ껄꧒녡꣓ꪺ압ꯈ걹ꖢ냝썄ꅁꖻ곣ꡳ껚뻚뱶암맘쁉꒽ꕱꪺꭏ돦ꚭ듁ꖢ껄꒧ꭏꓡꙝ꿀ꅂꭏ돦ꙝ꿀ꅂ껖ꭏꙝ꿀ꅂ띾냈ꑈ귻ꙝ꿀뭐롧샙ꙝ꿀ꅁ륂ꗎ룪껆뇄쑱덎꣓Ꝁ륷듺뭐꓀쏾ꅁ뙩ꛓꯘꗟꞹ뻣ꪺ뗻꛴볒ꮬꣃꓱ룻꣤륷듺껄ꩇꅁ듁꿠ꚳ껄ꙡ륷듺ꭏ돦ꚭ듁ꖢ껄ꪺ볧Ꙣꭏꓡꅁꕈ꣏ꭏ쁉롧샧꿠냷끷맯꣣ꚳ낪ꯗꭏ돦ꖢ껄궷쁉꒧ꭏꓡꅁ뺨ꚭ뇄꣺뽮랥ꪺꮴ곹ꭏꗾꑵꝀꅁꕈ쇗ꝋ띾냈ꭾ뷨ꪺ쑾쓲둣꓆ꛓ뷄삻꣬꒽ꕱꪺ샧륂ꅃ 2. ꓥ쑭Ꙟ압 Norman (1960) 끷맯뱶암ꭏ돦ꚭ듁ꖢ껄ꙝ꿀꒧곣ꡳ꒤땯뉻ꅁꙢꭏꓡ꽓꧊ꓨ궱ꅁꗕ믢뚥꿅ꅂꖮ꣠꒤ꛑ꙾ꑈꅂ귬꒽ꕱꭏꓡꪺ덑ꭏ쁉ꑈ결ꭏ돦ꚭ듁ꖢ껄ꕩ꿠꧊룻ꝃ꒧롳엩ꅆꙢꭏ돦꽓꧊ꓨ궱ꅁ뇄꙾쎺ꭏ뙏ꓨꚡꅂꛛ냊맔쎺ꅂ셾룪ꚩ쎺ꅂꭏ뙏룻낪ꪺꭏ돦ꅁ꣤ꖢ껄ꕩ꿠꧊돌ꝃꅆ낣ꚹ꒧ꕾꅁ돌궫굮ꪺ걏띾냈ꑈ귻ꪺ녍띾ꓴ럇꒣ꢬ깥꧶뻉교낪썂ꭏ돦ꪺꖢ껄ꅁꛓ띾냈ꑈ귻꙰ꩇ꿠Ꙣꭏ돦ꗍ껄ꯡꪺꕢ꙾꣬ꑀ꙾꒺꯹쓲ꙡ뒣꣑곛쏶ꪺꩁ냈ꅁ녎ꕩꑪꑪꙡ궰ꝃꭏ돦ꪺꚭ듁ꖢ껄뉶ꅃ놩꣤뛈뇄ꛦꖭꞡ볆ꪺꓱ룻뭐뮡ꧺꅁꣃꡓꚳ륂ꗎ뙩뚥ꪺ닎군ꓨꩫ꒩ꕈ샋ꥷ뭐꓀꩒ꅁꙝꚹ때ꩫ녯ꪾꭏ돦ꚭ듁ꖢ껄뭐ꙝ꿀뚡ꪺ곛쏶땻ꯗꅃ 과냪맘쁉ꛦ빐곣ꡳꣳ라ꅝLife Insurance Market and Research Associationꅁ슲뫙LIMRAꅞ꧳ꭏ돦ꗍ껄ꯡꪺꫬ꙾ꯗꅁꕈꕢ꙾뷕걤ꑀꚸꪺꓨꚡ곣ꡳꭏ돦ꪺꖢ껄뉶ꅁꣃꕂ땯ꛦꚳ쏶ꖢ껄뉶ꪺꕚꪫ돸ꝩꅁ낣ꑆ끷맯꒣Ꙑ뫘쏾꒧쁉뫘ꅁ꣒꙰꙾ꭏ쁉ꅂ롕꿠맘쁉ꅂꖢ꿠쁉ꑀ꿫맘쁉떥ꅁꚳꥷ듁ꪺꭏ돦쑾쓲뉶곣ꡳ돸ꝩꕾꅁꣃ맯꧳과냪맒ꕾꙡ냏꒧라귻꒽ꕱꚳꞹ뻣ꪺ곛쏶꓀꩒ꅁ껚뻚덯곣ꡳ돸ꝩꕘꅁ뱶암ꭏ돦ꚭ듁ꖢ껄ꪺꙝ꿀ꚳꭏꓡꪺ꧊ꝏꅂ꙾쓖ꅂ녂ꯃꪬꩰꅂ슾띾ꅂꫀ라ꙡꛬꅂ쎺뙏ꓨꚡꅂ걏ꝟ결귬ꭏꓡꅂꢭ엩ꪬꩰꅂꛦ빐덱룴ꅂ띾냈ꑈ귻꧊ꝏꅂ걏ꝟ결ꥴ꣠ꭏ돦ꅝorphan policyꅞꅂꭏꓡ둎띾ꪬꩰꅂꭏ뙏굴뻡꿠ꑏ떥ꙝ꿀ꅁꕄ굮ꕩ쉫쏾결ꭏꓡꅂꭏ돦ꅂ띾냈ꑈ귻ꅂ껖ꭏ롧샙꒭ꑪꙝ꿀ꅝCarter, 1995ꅂ Kelly, 1996ꅂSondergeld, 1997ꅂ Purushotham, 2001ꅞꅃꚹꕾꅁ결꿠ꚳ껄ꙡ룑ꡍ낾낪ꪺꭏ돦ꚭ듁ꖢ껄뉶꒧냝썄ꅁ과냪맘쁉ꛦ빐곣ꡳꣳ라꧳1982꙾땯깩ꕘ륷꛴ꭏ돦꯹쓲꧊땻ꯗ꒧ꑵ꣣ꇐ쑾쓲뉶볆ꅝpersistency raterꅞꅁꗎꕈ륷듺ꭏ돦ꗍ껄ꯡ닄ꑇꑑ꒭귓ꓫ껉꒴쑾쓲ꚳ껄꒧ꕩ꿠꧊ꅁ룓볆껚뻚덑ꭏ쁉ꑈ꽓꧊ꅂꭏ돦꽓꧊뭐꣤ꕌ꽓꧊ꕛꕈ뗻꓀ꛓ륷듺룓ꭏ돦꒧쑾쓲뉶볆ꅁ낵결ꭏ쁉꒽ꕱ띾냈ꑈ귻걏ꝟ뇄꣺ꟳ뽮랥ꪺꮴ곹껄ꑏ
344맘쁉ꭏ돦ꚭ듁ꖢ껄꒧륷듺 뫻꯹꒧ꛦ냊ꅁ룓볆ꣃꡓꚳꛒ뱻꣬껖ꭏꅂ띾냈ꑈ귻뭐롧샙떥ꙝ꿀맯ꭏ돦껄ꑏꪺ뱶암ꅃ 쏶꧳롧샙꽓꧊맯ꭏ돦ꚭ듁ꖢ껄뱶암ꪺ곣ꡳꅁOutreville (1990) ꕈ1966꣬1979꙾뚡과냪뭐ꕛ꺳ꑪꪺ셠엩롧샙닎군룪껆ꅁ맪쏒땯뉻ꝑ뉶뭐ꭏ돦ꚭ듁ꖢ껄뉶꒧뚡ꣃ때엣뗛ꪺ곛쏶ꅁCox et al. (1992) ꭨꕈ1980ꛜ1990꙾결볋ꖻ듁뚡ꅁ맪쏒땯뉻ꭏ돦륷ꥷꝑ뉶뭐ꖫ돵ꝑ뉶꒧깴늧띕ꑪꭨꖢ껄뉶녎라뱗ꕛꅁ꣢ꝥ뉻엣뗛ꪺꖿꙖ쏶ꭙꅁꙝ결ꝑ뉶꧳룓듁뚡엜냊룻ꑪꅁ걇Tsai et al. (2002) 뭻결ꕩ꿠걏볋ꖻ듁뚡꓀꩒ꓨꩫ꒧깴늧ꅁ교꣏ꑗ굺꣢뵧ꓥ쑭꒧맪쏒떲ꩇꚳ꧒꒣Ꙑꅃ꣤ꕈꙀ뻣ꙘꙖ뙱ꛛ끪쉫볒ꚡ꓀꩒1959꣬1995꙾뚡ꪺ룪껆땯뉻ꅁ둎듁쇍뛕ꛓꢥꅁꭏ돦ꚭ듁ꖢ껄뉶라쁈뗛ꖫ돵ꝑ뉶ꪺ뮼뱗ꛓ뒣낪ꅆ둎땵듁ꛓꢥꅁ럭듁뭐ꭥ꣢듁ꖫ돵ꝑ뉶ꪺ깴띕ꑪ껉ꅁꭨ럭듁ꪺꭏ돦ꚭ듁ꖢ껄뉶라띕ꝃꅃ뻣엩ꛓꢥꅁ럭륷ꥷꝑ뉶ꑪ꧳ꖫ돵ꝑ뉶껉ꅁꭏ돦ꚭ듁ꖢ껄뉶라궰ꝃꅁꕂ럭륷ꥷꝑ뉶뭐ꖫ돵ꝑ뉶꒧깴띕ꑪ껉ꅁ꣤맯ꚭ듁ꖢ껄뉶꒧뱶암껄ꩇ띕엣뗛ꅃ 냪ꑈ맘ꭏ쁉Ꙑ띾꒽라ꡃ꙾꧒ꕘꪩꪺ뭏왗맘쁉띾귓ꑈ맘쁉꙾ꯗ롧엧ꚺꑠ뉶뭐룑곹ꖢ껄뉶곣ꡳ돸ꝩ꒤ꅁꭙꕈ맘쁉꒽ꕱ꧒ꗓ돸ꪺ맪믚룪껆ꅁ꓀ꝏꕈꭏ썂뭐ꗳ볆ꅁ뇄덶돦ꪺꓨꚡ껖뫢ꡄ녯ꙕ꙾ꯗꭏ돦ꖢ껄뉶ꅁꣃ꧊ꝏꅂ꙾쓖ꅂ엩샋ꝏ뭐ꭏ쁉꙾ꯗ군뫢ꙕ왛맮꙾ꯗ냪꙾듁귓ꑈ맘쁉ꭏ돦ꖢ껄뉶꒧ꓴ럇ꅁ맯꧳ꭏ뙏꒧굱ꥷ꣣ꚳ곛럭궫굮ꪺ냑ꛒ믹귈ꅃ곣ꡳ돸ꝩꕘꅁ냪Ꙣ꧊ꝏꪺ깴늧ꑗꅁꑫ꧊ꪺꭏ돦ꖢ껄뉶낪꧳ꡫ꧊ꅆ때엩샋ꪺꭏ돦ꖢ껄뉶룻낪ꅆꙢ꙾쓖ꓨ궱ꅁꖢ껄뉶쁈뗛ꭏꓡꪺ꙾쓖뱗ꕛꛓ궰ꝃꅆꭏ돦꙾ꯗꛓꢥꅁꕈ닄ꑀ뭐닄ꑇꭏ돦꙾ꯗꪺꖢ껄뉶돌낪뭐ꚸ낪ꅁ뙗륌꣢꙾ꪺꭏ돦ꅁ꣤ꖢ껄뉶ꭨ쇍꧳쎭ꥷꅃ놩꣤ꕘꪩꪺ돸ꝩ꒤꧒뇄ꗎ꒧곣ꡳꓨꩫ뛈궭꧳ꖭꞡ볆꒧ꓱ룻ꅁꣃꖼ뙩ꑀꡂ꓀꩒ꙕ뚵엜볆뭐ꭏ돦ꖢ껄뉶뚡걏ꝟ꣣ꚳ엣뗛ꪺ곛쏶ꅃꚹꕾꅁ꣤꧒뇄ꗎꪺ엜볆ꭙꕈꭏꓡ뭐껖ꭏ꒧꽓꧊결ꕄꅁꣃꖼ둎ꭏ돦ꖻꢭꅂ띾냈ꑈ귻꽓꧊꒧ꙝ꿀ꕛꕈ꓀꩒ꅃ 램ꥶ (1992) 뭐ꩌ쒣ꩆ (1998a) ꭨ륂ꗎ엞뿨꽓끪쉫볒ꮬꅁꝑꗎꕸꕟꖫꑈ맘ꭏ쁉Ꙑ띾꒽라꧒뭳Ꝁꪺꕸ왗맘쁉띾닄ꑔꙞ롧엧ꗍꥒꫭ꒧룪껆ꅁ놴끑뱶암냪꙾듁귓ꑈ맘쁉ꭏ돦닄ꑑꑔ귓ꓫꖢ껄꒧궫굮엜볆ꅁ덑ꭏ쁉ꑈ꽓꧊ꅂꭏ돦꽓꧊껖ꭏꙝ꿀ꕛꕈ꓀꩒ꅁ꣤곣ꡳ떲ꩇ땯뉻ꅁꭏ돦ꚭ듁ꖢ껄뭐ꭏꓡ꧊ꝏꅂꭏ꙾쓖ꅂ쎺뙏ꚸ볆ꅂꭏ쁉썂ꅂ꙾쎺꓆ꭏ뙏ꅂ쎺뙏듁뚡ꅂ쎺뙏ꓨꚡꅂ엩샋ꝏ뭐볐럇엩꒧ꭏꓡꅂ껖ꭏꭏ돦꽓꧊꒧ꙝ꿀꣣ꚳ엣뗛ꪺ곛쏶ꅁꓥ꒤ꣃꡓꚳ뒣룓볒ꮬ꒧륷듺꿠ꑏꅃꑗ굺ꓥ쑭끷맯냪맘쁉ꭏ돦ꚭ듁ꖢ껄꒧ꙝ꿀ꪺ곣ꡳ꒤ꅁ곒ꙝ룪껆꒧궭꣮ꅁ때ꩫ뙩ꑀꡂ꓀꩒띾냈ꑈ귻ꙝ꿀맯ꚭ듁ꖢ껄꒧뱶암ꅃ땍ꛓꅁ껚뻚과냪맘쁉ꛦ빐곣ꡳꣳ라ꪺ돸ꝩꅁ룓ꣳ라ꑷꧺ뵔ꙡꕘ띾냈ꑈ귻꒧ꩁ냈ꭾ뷨맯ꭏꓡ걏ꝟ룑곹ꚳꭄ녠엣뗛ꪺ뱶암ꅃꩌ쒣ꩆ
궷쁉뫞뉺뻇돸 345닄꒭ꣷ 닄ꑔ듁 2003꙾11ꓫ (1998b) ꕴ뻚빐냢띾냈ꑈ귻꒧꽓꧊ꅁꕈ뷆끪쉫볒ꮬ꓀꩒땯뉻ꅁꭏ돦쑾쓲뉶뭐꒽ꕱ띾냈ꑈ귻ꙁ땮뿽뉶띾냈ꑈ귻땮뿽ꛒ룕껦뉶ꚳ엣뗛ꪺ곛쏶ꅁꗑꚹꕩꪾ띾냈ꑈ귻꒧걹냊꧊뭐녍띾ꓴ럇맯ꭏ돦ꚭ듁ꖢ껄ꚳ엣뗛꒧뱶암ꅁ땍ꛓ룓볒ꮬ걏ꕈꙕ꒽ꕱ뚡꒧뻣엩띾냈ꑈ귻꽓꧊ꪺ깴늧꣓뿅뙱ꗾ꒽ꕱ꙾ꯗꖢ껄뉶ꪺꓴ럇ꅁꣃꡓꚳꙐ껉ꕛꕈꛒ뙱덑ꭏ쁉ꑈꅂꭏ돦ꅂ껖ꭏ띾냈ꑈ귻ꖻꢭꪺ꽓꧊ꅁꙝꚹ때ꩫ끷맯귓ꝏꭏꓡꪺꭏ돦ꖢ껄꒧ꕩ꿠꧊ꕛꕈꝐ쉟ꅃ 3. 곣ꡳ룪껆뭐ꓨꩫ ꖻ곣ꡳꪺ볋ꖻ룪껆결냪꒺걙ꑀ맘쁉꒽ꕱ87꙾ꯗ띳ꗍ껄꒧꙾듁맘쁉ꮴ곹ꅁꙀ군12,140떧ꅁꣃ녱ꗾ뎡ꪺ볋ꖻ꒤쁈뻷ꧢꕘ3,035떧룪껆낵결볒ꮬꪺ륷듺꿠ꑏ꒧듺룕룪껆ꅃ궺ꗽꅁ냵ꛦ엞뿨꽓끪쉫ꪺ덮엩결SPSS ꪩꅁ뇄ꗎꙖꯡ뇸ꗳꚡ덶ꡂ끪쉫뽺뿯ꕘ엣뗛ꪺꛛ엜볆ꅃ꣤ꚸꅁ쏾꾫롧뫴룴ꪺ끖뵭뭐듺룕ꭨ꣏ꗎSTATISTICA Neural Network ꪩꅁꚳ쏶뫴룴볒ꮬꪺꯘꗟ뭐뿩ꑊ엜볆ꪺ뿯꣺ꭨꚳꑔ뫘꒣Ꙑꪺꓨꚡꅁ꓀ꝏ결뇄ꗎꗾ뎡ꪺ륷듺엜볆ꅂꝑꗎ엞뿨꽓끪쉫꧒뽺뿯ꕘꪺ륷듺엜볆꣏ꗎ덮엩꒤ꪺꛛ냊뫴룴덝군(automatic network designer)꒧ꕜ꿠ꅁꓱ룻ꑔ뫘떲ꩇꯡꙁꕘ돌꣎ꪺ뫴룴볒ꮬꅃ 엜볆뮡ꧺ ꖻ곣ꡳꭙꕈꭏ돦ꗳ볆결냲슦ꅁꣃ껚뻚끝걆뎡ꭏ쁉ꕱ꧒륻ꖬ꒧띾냈롧샧ꛛꯟ럇ꭨ멛엳엩꣮떽군릺꒤ꪺꥷ롱ꅁꚭ듁ꭏ돦ꖢ껄ꭙꭏ돦ꗍ껄ꯡ닄ꑇꑑ꒭귓ꓫ꒺룑곹ꅝsurrenderꅞꅂ낱껄ꅝterminationꅞꖢ껄ꅝlapseꅞ꒧ꭏ돦ꅁ꒣ꕝ걁껄ꑏ닗ꓮꅝ꙰ꚺꑠ떹ꕉꅞ뭐ꮴ곹멍빐꒧ꭏ돦ꅁꣃ뇆낣쒻쎺뭐ꑀ꙾듁ꥷ듁맘쁉ꭏ돦ꅃ결꿠ꕒ꓀ꑆ룑맘쁉꒽ꕱꚭ듁ꭏ돦ꖢ껄뉶륌낪꒧냝썄ꅁꙝꚹꖻ곣ꡳꕈ닄ꑇꑑ꒭귓ꓫ걏ꝟꖢ껄결럇ꭨ엜볆ꅁ럭땍ꗧꕝꝴ닄ꑑꑔ귓ꓫꖢ껄꒧ꭏ돦ꅃꕴ껚뻚꧒냑ꛒꓥ쑭ꪺ떲뷗뭐룪껆꣺녯꒧궭꣮ꅁꖻ곣ꡳꕈꭏꓡꅂꭏ돦ꅂ껖ꭏ뭐띾냈ꑈ귻떥ꕼ뫘ꙝ꿀결륷듺엜볆ꅁ엜볆ꥷ롱뭐꒺깥뮡ꧺ꙰ꫭ1꧒ꗜꅃ쇶땍ꖫ돵ꝑ뉶ꪺ엜꓆ꅂꭏꓡ꧒녯엜꓆뭐둎띾ꪬꩰ떥ꗧ결뱶암ꭏ돦ꚭ듁ꖢ껄꒧궫굮ꙝ꿀ꅝCox et al., 1992ꅆCarter, 1995ꅆTsai et al., 2002ꅞꅁꙝ꧒곣ꡳꪺ꒽ꕱ꒧ꯈꓡ룪껆깷꒤ꣃ때ꭏꓡꪺ꧒녯뭐둎띾ꪬꩰ떥끏뿽ꅁ맪ꑗꅁ냪꒺맘쁉꒽ꕱ둘ꕇ뎣ꖼꯘꗟꚹꓨ궱ꪺ룪껆ꅁꙝꚹꫬꡂ뇆낣롧샙ꙝ꿀꒧ꛒ뙱ꅃꚹꕾꅁꖻ곣ꡳꭙꕈ돦ꑀ꒽ꕱ결곣ꡳ볋ꖻꅁ걇때ꩫꓱ룻꒽ꕱ뚡ꪺ깴늧ꅁ꣤떲ꩇꕩ삳ꗎ꣬꣤ꕌꗴ꛳꒽ꕱꅃ 쏾꾫롧뫴룴Ꙣ끖뵭ꭥ덱녠뚷녎뿩ꕘꅂꑊ엜볆걍깧꣬Ꙙ뉺ꪺ냏뚡료ꅁ뫙결엜볆ꓘꯗ꓆ꅝscalingꅞꅁ껚뻚둞ꓥꖿꅝ2001ꅞ꧒굺ꅁ쏾꾫롧뫴룴ꪺ뿩ꑊ뭐뿩ꕘ귈꒶꧳0꣬1뚡ꪺ껄ꩇ돌꣎ꅃꙝꚹꙢ꣏ꗎꭥꖲ뚷ꗽ녎엜볆뙩ꛦ신뒫ꅁ
346맘쁉ꭏ돦ꚭ듁ꖢ껄꒧륷듺 ꛓ꣤ꓨꚡꙝ룪껆ꪺ쏾ꮬ꒣Ꙑꛓꚳ꧒깴늧ꅁꖻ곣ꡳꕈ뻷뉶맯걍ꩫ덂뉺ꓱ뉶ꓘꯗꅝratio scaleꅞꪺ엜볆ꅁꕈ냏뚡맯걍ꩫ덂뉺냏뚡ꓘꯗꅝinterval scaleꅞ뭐뚶Ꟈꓘꯗꅝordinal scaleꅞꪺ엜볆ꅁꛜ꧳ꙗꗘꓘꯗꅝnominal scaleꅞꭨꕈ뗪샀엜볆뙩ꛦ신뒫ꅁ꙰ꫭ1꧒ꗜꅃ꣤꒤ꚳ쏶ꥴ꣠ꭏ돦ꪺꥷ롱결ꭏ돦ꖼꖢ껄ꭥꅁ귬ꩁ냈ꪺ띾냈ꑈ귻ꝙꑷ싷슾뫙꒧ꅃ ꫭ1₹瞴鉶?욻ꆩ請僂ꮤ? 엜볆 신뒫륷 듺 엜 볆 뮡 ꧺ 쏾ꝏ ꓨꚡ1.덑ꭏ쁉ꑈ꧊ꝏ(x)1ꫭꡫ꧊ꅁ0ꫭꑫ꧊ꅃ 뗪샀12덑ꭏ쁉ꑈ꒧ꭏ꙾쓖(x) 군뫢ꭏ껉ꪺ맪믚꙾쓖ꅃ 냏뚡23덑ꭏ쁉ꑈ꒧녂ꯃ(x) 1ꫭꑷ녂ꅁ0ꫭꖼ녂ꅃ 뗪샀34덑ꭏ쁉ꑈ꒧슾띾쏾ꝏ(x) ꕈ슾띾떥꿅1-6쏾ꫭꗜꅃ 냏뚡ꭏ 4ꓡ 5.굮ꭏꑈ꙾쓖(x) 군뫢ꭏ껉ꪺ맪믚꙾쓖ꅃ 냏뚡5꽓 6.굮ꭏꑈ꧊ꝏ(x)1ꫭꡫ꧊ꅁ0ꫭꑫ꧊ꅃ 뗪샀6꧊ ꕈ6귓뗪샀엜볆꓀ꝏꫭꑬꑫꅂꕓꥪ7.뭐덑ꭏ쁉ꑈ꒧쏶ꭙ(xꇐx) ꥦꅂ끴낸ꅂꖻꑈꅂꓷꗀꅂ깡ꑈ뭐꣤ꕌ뗪샀712쏶ꭙꅃ 8.걏ꝟ결귬ꭏꓡ(x) 1ꫭ걏ꅁ0ꫭꝟꅃ 뗪샀13ꝑꗎ3귓뗪샀엜볆꓀ꝏꫭꗜ꙾ꅂꕢ9.쎺뙏ꝏ(xꇐx) 뗪샀1416꙾ꅂꥵ뭐ꓫ쎺떥ꓨꚡꅃ ꝑꗎ2귓뗪샀엜볆꓀ꝏꫭꗜꛛ냊신10.쎺뙏ꓨꚡ(xꇐx) 뗪샀1718녢ꅂ뚰쎺뭐ꚬ뙏떥ꓨꚡꅃ ꭏ 11.쎺뙏꙾듁(x) ꓀결6ꅂ10ꅂ15ꅂ20꙾ꅃ 냏뚡19돦 12.ꕄ곹ꭏ쁉썂(x) ꣺ꛛ땍맯볆ꅃ 뻷뉶꽓 20꧊ 13.꙾쎺꓆ꕄ곹ꭏ뙏(x) ꣺ꛛ땍맯볆ꅃ 뻷뉶2114.ꚳ때곹(x) 1ꫭꚳꅁ0ꫭꝟꅃ 뗪샀2215.꙾쎺곹ꭏ뙏(x) ꣺ꛛ땍맯볆ꅃ 뻷뉶2316.곹셠ꭏ뙏ꓱ꣒(x) 곹ꭏ뙏/셠ꭏ뙏ꅃ 뻷뉶24껖 17.ꚳ때엩샋(x) 1ꫭꚳꅁ0ꫭ때ꅃ 뗪샀25ꭏ 꽓 18.걏ꝟ볐럇엩(x) 1ꫭꚳꅁ0ꫭ때ꅃ 뗪샀26꧊ 19.띾냈ꑈ귻꧊ꝏ(x) 1ꫭꡫ꧊ꅁ0ꫭꑫ꧊ꅃ 뗪샀27띾 20.띾냈ꑈ귻꙾쓖(x) ꭏ돦ꗍ껄껉띾냈ꑈ귻ꪺ꙾쓖ꅃ 냏뚡28냈 ꑈ21.띾냈ꑈ귻ꩁ냈꙾룪(x) ꣬슾ꛜꭏ돦ꗍ껄꒧듁뚡 냏뚡 29귻 ꕈ1-6꓀ꝏꫭꗜ냪ꑰꅂ냪꒤ꅂ낪꒤ꅂ22.띾냈ꑈ귻뇐땻ꯗ(x) 냏뚡30꽓 녍곬ꅂꑪ뻇ꅂ곣ꡳ꧒ꕈꑗ떥뇐땻ꯗꅃ ꧊ 23.걏ꝟ결ꥴ꣠ꭏ돦(x) 1ꫭ걏ꅁ0ꫭꝟꅃ 뗪샀31 곣ꡳꓨꩫ
궷쁉뫞뉺뻇돸 347닄꒭ꣷ 닄ꑔ듁 2003꙾11ꓫ Ꙣ뛇닎ꪺꓨꩫ꒤ꅁ엞뿨꽓끪쉫ꪺ삳ꗎꙝ꒣믝굮릳냏ꝏ꓀꩒ꪺꙨ엜뙱녠멁꒧낲덝ꅁ걇Ꙣꑇ꓀쏾ꅝbinary or dichotomousꅞ냝썄ꪺ곣ꡳ룻뱳꩸ꙡ덑뇄ꗎꅁ꣒꙰ꭏ돦ꖢ껄륷듺ꅝ램ꥶꅁ1992ꅆꩌ쒣ꩆꅁ1998aꅞꅂ꽽늣륷듺ꅝꝦ맅곕ꅁ2000ꅞ떥ꅃ엞뿨꽓끪쉫ꪺ냲ꖻꚡ뭐ꑀ꿫뵵꧊끪쉫ꣃ때ꯜꑪ꒣Ꙑꅁ놩럇ꭨ엜볆꒣ꙁ꙰뵵꧊끪쉫결덳쓲꧊엜볆ꅁꛓ걏ꕈꑇ꓀쏾엜볆멁ꕘ뉻ꅁꙝꚹꖲ뚷덺륌엞뿨꽓끪쉫꣧볆ꪺ신뒫꣏꣤뭐륷듺엜볆ꝥ뉻뵵꧊ꪺ쏶ꭙꅃꕏT럇ꭨ엜볆y = 01꓀ꝏꫭꗜꭏ돦ꚳ껄ꖢ껄ꅆβ=[β,β,Λβ]ꫭꗜ끪0131T쉫ꭙ볆Ꙗ뙱ꅆx=[1,x,xΛx]ꫭꗜ륷듺엜볆Ꙗ뙱ꅁ꣤볒ꮬ꙰ꑕ꧒ꗜꅃ 1231g(x)eπ(x)=(1)g(x)1+e31T꣤꒤g(x)=βx=β+βxꅃ뒫ꕹ룜뮡ꅁπ(x)ꝙꫭꗜy = 1ꪺ땯ꗍ뻷뉶ꅁ∑0iii=1귌뫙꒧결엞뿨꽓끪쉫꣧볆ꅝlogistic regression functionꅞꅃ굙녎ꚹ꣧볆Ꝁꕈꑕ신뒫ꅝHosmer and Lemeshow, 1989ꅞꅇln{π(x)/[1−π(x)]}ꅁꭨꕩ녯ꑀ귓뵵꧊끪쉫볒ꮬꅃ뙩ꑀꡂꛓꢥꅁꙝπ(x)=E(y=1|)결xꅁ걇뇄ꗎꚹ볒ꮬ껉ꖲ뚷ꡍꥷ셻곉귈ꅝcut-off valueꅞꅁꕈꝀ결ꡍꥷ걏ꝟ쓝ꖢ껄ꭏ돦ꪺ꓀쏾볐럇ꅁ덱녠꣺돌ꭏꙵꪺ셻곉귈결ꅁ럭ꭏ돦ꚭ듁ꖢ껄뉶뭐ꑇ꓀꒧ꑀꚳ엣뗛ꪺ깴늧껉ꅁꭨ뚷ꝑꗎ엧ꭥ뻷뉶ꅝprior probabilityꅞꝀ결셻곉귈ꅃ 귋뛇뮼쏾꾫롧뫴룴ꅝback-propagation network, 슲뫙BPNꅞꭨ걏쏾꾫롧뫴룴꒤삳ꗎ돌결뱳꩸ꪺ볒ꮬꅁ녠덑ꗎ꣓룑ꡍꑇ꓀쏾냝썄ꅁ꙰ꖢ뉍쁶꿠ꑏ꒧륷듺ꅝBrockett et. al., 1994ꅆ걟ꭔꅁ1994ꅆ남꒤얻ꅁ1996ꅞꅂ믈ꛦ뇂ꭈꗸ띾꒧륈곹궷쁉륷듺ꅝ뎯쁁ꟸꅂ덜덱ꙷꅂꩌ붯꾳(1996)떥ꅃ꣤뭐Ꙩ뱨꣧볆덳떲뫴룴ꅝmultiplayer functional-link network, 슲뫙MFLNꅞ곒쓝꧳Ꙩ뱨띐ꪾ뻷ꅝmulti-layer preceptronsꅞ꒤ꪺ뫊럾ꚡ뻇닟ꅝsupervised learningꅞ뫴룴볒ꮬꅁꙨ뱨꣧볆덳떲뫴룴ꕈꑔ귓륂뫢꒸꣓ꫭ륆ꑀ귓뿩ꑊ엜볆ꅁ꓀ꝏ결ꖿ덗꓆ꅂ맯볆꓆볆꓆신뒫ꅁꗑ꧳맯볆꓆륂뫢꒸맯엜볆ꪺꝃ귈냬뎡ꗷ룻뇓빕ꅁ볆꓆륂뫢꒸맯엜볆ꪺ낪귈냬뎡ꗷ룻뇓빕ꅁꙝꚹꙨ뱨꣧볆덳떲뫴룴녠꿠ꚳ룻낪ꪺ뫫럇ꯗꅝ뢭꧉ꚨꅁ1997ꅞꅃ 귋뛇뮼쏾꾫롧뫴룴뭐Ꙩ뱨꣧볆덳떲뫴룴꧒녯ꪺ뿩ꕘ귈뭐끖뵭뵤꣒ꪺꗘ볐귈꒬곛ꓱ룻ꕩꡄ녯뫴룴뿹뭾ꅁꣃꝑꗎ돌끾궰ꩫꅝgradient steepest descent methodꅞ녎ꚹ뿹뭾Ꝁ결귗ꖿ덳떲꒤ꪺꕛ앶귈뭐믖귈ꪺ뻚ꅁ뙩ꛓ녱끖뵭뵤꣒꒤ꯘꗟ돌꣎ꪺ륷듺볒ꮬꅝFausett, 1994ꅞꅃ꿷둎ꖻ곣ꡳ꧒삳ꗎꪺ귋뛇뮼쏾꾫롧뫴룴볒ꮬ뮡ꧺ꙰ꑕꅇ
348맘쁉ꭏ돦ꚭ듁ꖢ껄꒧륷듺 nh=f(wx−θ) , j=1,2,Λ,m (2) ∑jijiji=1mˆy=f(ch−λ)(3)∑jjj=1꣤꒤ꅇnꅂm꓀ꝏꫭꗜ뿩ꑊ뱨뭐쇴싃뱨ꪺ륂뫢꒸볆ꗘꅆ wꫭꗜ닄i귓뿩ꑊ엜볆뭐닄j귓쇴싃륂뫢꒸꒧뚡ꪺ앶볆귈ꅆ ijcꫭꗜ닄j귓쇴싃륂뫢꒸뭐뿩ꕘ륂뫢꒸꒧뚡ꪺ앶볆귈ꅆ jθꫭꗜ닄j귓쇴싃륂뫢꒸꒧믖귈ꅆ jλꫭꗜ뿩ꕘ륂뫢꒸꒧믖귈ꅆ xꫭꗜ닄i귓뿩ꑊ엜볆ꅁi=1,2,Λ,nꅆ ihꫭꗜ닄j귓쇴싃륂뫢꒸꒧뿩ꕘ귈ꅆ jˆyꫭꗜ뿩ꕘ륂뫢꒸꒧뿩ꕘ귈ꅁꗧ결뫴룴볒ꮬꪺ듁뇦뿩ꕘ귈ꅆ −η−1f(⋅)ꫭꗜꝀꗎ꣧볆ꅁꖻ곣ꡳꕈ엞뿨꽓꣧볆ꅇf(η)=(1+e)Ꝁ결쇴싃뱨뭐뿩ꕘ뱨륂뫢꒸꒧Ꝁꗎ꣧볆ꅝHaykin, 1999ꅞꅃ ꝑꗎꙨ뱨띐ꪾ뻷꣓룑ꡍ냝썄껉ꅁ궺ꗽꖲ뚷룑ꡍꪺ걏룓ꗎ둘뱨ꪺ걛멣ꅈꕈꡃꑀ뱨ꪺ륂뫢꒸ꪺ볆ꗘ걏Ꙩꓖꅈ뉺뷗ꑗꅁꙨ뱨띐ꪾ뻷ꕵ믝굮꣢뱨걛멣ꅁꙁꕛꑗ쇴싃뱨ꪺ륂뫢꒸냷Ꙩꪺ룜ꅁꙨ뱨띐ꪾ뻷ꪺ뿩ꕘꭋꕩ륇꫱ꗴ띎덳쓲꣧볆ꅃ꓀꩒ꑈ귻껚뻚볆뻚엜볆꽓뷨꣓ꡍꥷ뿩ꑊꅂ뿩ꕘ쇴싃뱨꒺륂뫢꒸ꪺ볆뙱ꅃꖻ곣ꡳ롧듺룕ꯡ땯뉻ꅁꕈꑀ귓쇴싃뱨ꪺ뫴룴걛멣ꅁ뻇닟둠샴(learning cycle)결1000ꅂ뻇닟덴뉶(learning rate)결ꅂ멄꧊ꙝꑬ(momentum)결껉ꚳ돌꣎ꪺ륷듺껄ꩇꅃ ꗑ꧳ꖻ곣ꡳꪺ볋ꖻ볆낪륆12,140떧룪껆ꅁꙝꚹꝙ꣏걏ꛊ꓀꒧ꑀꪺ엜꓆ꗧꚳꕩ꿠라덹ꚨꯜꑪꪺ뱶암ꅁ결ꓱ룻ꙕ뫘볒ꮬꪺ륷듺꿠ꑏꝟꚳ엣뗛ꪺ깴늧ꅁ귌ꕈ듺룕볋ꖻꪺ떲ꩇ꣓샋ꥷ볒ꮬꪺ륷듺꿠ꑏꅃ궺ꗽꕏp뭐p꓀ꝏꫭ12ꗜ꣢뫘ꓨꩫ꒧뭾Ꝑ뉶ꅁ걇뗪때낲덝결p뭐p때엣뗛ꪺ깴늧(ꝙHꅇ120p=p=p)ꅃꙝꚹꙢ뗪때낲덝결꽵ꪺ놡ꩰꑕꅁꝑꗎˆp=(x+x)(n+n)꛴121212군ꙀꙐꪺ뭾Ꝑ뉶pꅁ꣤꒤n뭐n꓀ꝏꫭꗜ꣢뫘ꓨꩫ꒧볋ꖻ볆ꅆx뭐x꓀1212ꝏꫭꗜ꣢뫘ꓨꩫ꒧뭾Ꝑ귓볆ꅃ걇ꕩ뇀녯샋ꥷ닎군뙱꙰ꑕꅇ ˆˆˆˆp−pp−p1212 (4) z==σˆˆˆˆˆp(1−p)n+p(1−p)n(pˆ−pˆ)1212꣤꒤ˆp=xn뭐pˆ=xn꓀ꝏꗎ꣓꛴군꽵맪ꪺ뭾Ꝑ뉶pꥍpꅁσˆ12111222(pˆ−pˆ)12
궷쁉뫞뉺뻇돸 349닄꒭ꣷ 닄ꑔ듁 2003꙾11ꓫ ꭨ결닎군뙱ˆˆp−pꪺ꛴군볐럇뭾ꅃ 볒ꮬ륷듺꿠ꑏ꒧뿅뙱 ꑀ꿫ꛓꢥꅁꙢ뗻꛴꓀쏾볒ꮬꪺ륷듺꿠ꑏ껉ꅁ녠ꕈ볒ꮬꪺ귓ꝏ뭐뻣엩ꖿ뵔뉶뭾Ꝑ뉶낪ꝃ뿅뙱륷듺ꪺ껄ꩇꅁꑇ꓀쏾냝썄맯셻곉귈ꪺ뿯ꥷ곛럭뇓띐ꅁ결Ꝋꩁ셻곉귈ꪺ뱶암ꅁZweig and Campbell(1993)ꕈ놵꣼뻞Ꝁ꽓뱸ꚱ뵵꧒ꕝꝴꪺ궱뽮(area under receiver operating characteristic curveꅆ슲뫙AUROC)ꪺꑪꑰ꣓ꓱ룻볒ꮬꙢ꒣Ꙑ셻곉귈ꑕ꒧꓀쏾떲ꩇꅁꣃ뗻꛴꓀쏾볒ꮬꪺ냏ꝏ껄ꩇꅃ껚뻚ꥷ롱ꅁꧺ뵔ꯗ(specificity)ꫭꗜꖿ뵔꓀쏾ꚳ껄ꭏ돦꒧ꓱ뉶ꅁ뇓띐ꯗ(sensitivity)ꫭꗜꖿ뵔꓀쏾ꖢ껄ꭏ돦꒧ꓱ뉶ꅁ덱녠AUROC귈꒶꧳0꣬1꒧뚡ꅁ띕ꑪꫭꗜ볒ꮬꪺ냏ꝏ껄ꩇ띕꣎ꅁ꣥ꮬꪺROCꚱ뵵꙰맏1꧒ꗜꅃ ROCꚱ뵵1 뇓띐ꯗ 0 11ꇐꧺ뵔ꯗ 맏1?ROCꚱ뵵맏 ꝑꗎAUROCꪺꑪꑰ꣓뿅뙱볒ꮬ냏ꝏ껄ꩇ껉ꅁꖲ뚷낲덝꓀쏾뿹뭾껉ꅁ1꣤ꮬI뿹뭾ꅝtype I errorꅞ뭐ꮬII뿹뭾ꅝtype II errorꅞ꧒덹ꚨꪺꚨꖻ뉶ꅝcost of ratioꅞ떥꧳1ꅁ꙰ꚹꪺ낲덝ꣃ꒣꓁맪믚ꅝAdams and Hand, 2000ꅞꅁꙝ결둎ꭏ돦ꖢ껄륷듺ꛓꢥꅁ녎라ꖢ껄ꪺꭏ돦쉫쏾결꒣라ꖢ껄ꭏ돦ꅝꝙꗇꑕꮬI뿹뭾ꅞ맯ꭏ쁉꒽ꕱꪺ룪뭐띾냈ꭾ뷨꒧뱶암ꅁ뮷룻녎라ꚳ껄ꪺꭏ돦쉫쏾결라ꖢ껄ꪺꭏ돦ꅝꝙꗇꑕꮬII뿹뭾ꅞ꒧뱶암꣓ꪺ쑙궫덜ꙨꅃꙝꚹBoritz and Kennedy (1995)ꅂLin (1996)뭐Boonyanunta and Zeephongsekul (2000)ꛒ뱻Ꙣ꒣Ꙑꪺꚨꖻ뉶ꑕꅁꕈ뭾Ꝑꚨꖻꅝmisclassification costꅞ꣓뿅뙱볒ꮬꙝ뭾Ꝑ꧒덹ꚨꪺ띬ꖢ땻ꯗꅁ뙩ꛓꓱ룻륷듺볒ꮬꪺ쁵Ꙉ꧊ꅃ뉺뷗ꑗꅁ럭뭾Ꝑꚨꖻ돌ꑰ껉ꅁꑇ꓀쏾볒ꮬꕩ껚뻚꽓ꥷꪺꚨꖻ뉶꣓ꡍꥷ돌꣎셻곉귈ꅁꚨꖻ뉶녠 1 ꚨꖻ뉶ꭙꙝ뭾Ꝑ늣ꗍ꒧ꮬI뿹뭾뭐ꮬII뿹뭾꧒뻉교ꪺ띬ꖢꚨꖻ꒧ꓱ뉶ꅃ셼꣒꣓뮡ꅁꚨꖻ뉶결10ꝙꫭꗜꙝꮬI 뿹뭾꧒덹ꚨꪺ띬ꖢꚨꖻ결ꮬII뿹뭾ꪺ10궿ꅃ
350맘쁉ꭏ돦ꚭ듁ꖢ껄꒧륷듺 ꙝ샴맒뫞뉺떥ꙝ꿀ꪺ엜꓆ꛓ엜ꅁꙝꚹꣃ때ꥷꪺ볆뻚ꕩ꣑냑ꛒꅃꖻ곣ꡳ녎껚뻚꧒ꯘꗟꪺ볒ꮬꅁꕈ볒샀ꪺꚨꖻ뉶꣓놴끑꣤엜꓆꧒덹ꚨꪺ뱶암ꅃ 껚뻚Boritz and Kennedyꪺꥷ롱ꅁꮬI뿹뭾걏맪믚결ꖢ껄ꭏ돦륷듺결ꚳ껄ꭏ돦꒧뭾Ꝑ뉶ꅁꮬII뿹뭾걏맪믚결ꚳ껄ꭏ돦륷듺결ꖢ껄ꭏ돦꒧뭾Ꝑ뉶ꅁ꣤뭾Ꝑꚨꖻꪺꥷ롱꙰ꑕꅇ 뭾Ꝑꚨꖻ=(A+B)/C(5)꣤꒤ꅇAꇗꮬI뿹뭾ꪺꛊ꓀ꓱ*ꖢ껄ꭏ돦ꪺꓱ뉶*ꚨꖻ뉶ꅆ BꇗꮬII뿹뭾ꪺꛊ꓀ꓱ*ꚳ껄ꭏ돦ꪺꓱ뉶ꅆ Cꇗꖢ껄ꭏ돦ꪺꓱ뉶*ꚨꖻ뉶ꇏꚳ껄ꭏ돦ꪺꓱ뉶ꅃ 뻣엩ꛓꢥꅁꖻ곣ꡳꗽꕈ볒ꮬꙢ끖뵭뭐듺룕볋ꖻꪺ뭾Ꝑ뉶꣓ꓱ룻ꅁꣃ샋ꥷ볒ꮬ꒧륷듺꿠ꑏ걏ꝟ꙳Ꙣ엣뗛ꪺ깴늧ꅁꙁꕈAUROC뇆낣셻곉귈깴늧맯볒ꮬ냏ꝏ껄ꩇꪺ뱶암ꅃꕴꕾꅁ껚뻚볒ꮬꪺ뭾Ꝑꚨꖻ꣓ꓱ룻엞뿨꽓끪쉫뭐쏾꾫롧뫴룴ꪺ쁵Ꙉ꧊ꅁꣃ샋ꥷ꣢걏ꝟ꙳Ꙣ엣뗛ꪺ깴늧ꅁ돌ꯡꕈ돌빁ꪺ볒ꮬ뮡ꧺ뱶암맘쁉ꭏ돦ꚭ듁ꖢ껄꒧궫굮ꙝ꿀ꅃ 4. 곣ꡳ떲ꩇ 뇔굺꧊닎군 녱ꫭ2ꕩꕈ땯뉻ꅁ꙾쎺ꕄ곹뭐꙾쎺곹ꭏ뙏ꪺ볐럇깴곒ꑪ꧳꣤ꖭꞡ볆ꅁꛓꕂꭏ쁉썂뭐곹ꭏ뙏ꓱ꣒ꑝꚳ쏾ꪺ놡ꩰꅁ엣ꗜꭏ돦뚡ꪺꭏ믙깴늧꧊믡ꑪꅃꕴ띾냈ꑈ귻뭐굮ꭏꑈꪺꖭꞡ꙾쓖꓀ꝏ결뭐랳ꅁ녯ꪾ꣤띾냈ꑈ귻ꪺꛦ빐맯뙈ꕈ꙾쓖곛꫱ꪺꯈꓡ결ꕄꅆ낣ꚹ꒧ꕾꅁ룓꒽ꕱꪺ띾냈ꑈ귻꒧ꖭꞡꩁ냈꙾룪결꙾ꅁꕩ꿠걏뻉교ꭏ돦ꚭ듁ꖢ껄ꪺ귬ꙝ꒧ꑀꅃ ꕈ뗪샀엜볆ꪺ꓀ꝇ놡ꩰ꣓곝ꅁ꙰ꫭ3ꅁ꣤닄ꑇꑑ꒭귓ꓫꪺꭏ돦ꖢ껄뉶결%ꅁ낪꧳뻣엩맘쁉ꖫ돵ꪺ32%ꅃꛜ꧳ꭏꓡ쎺뙏ꓨꚡꭨꕈꚬ뙏귻결ꕄꅁ%ꅆ쎺뙏꙾듁ꭨꕈ20꙾듁결ꕄꅁ%ꅁꕂꕈ꙾쎺%돌Ꙩꅃ쏶꧳ꭏꓡꓨ궱ꅁꕈꛛꑶꓷꗀ결굮ꭏꑈꙨꅁ%ꅃ껖ꭏꙝ꿀ꓨ궱ꅁ볋ꖻ꒤ꕄ굮결볐럇엩ꕂ때엩샋ꅁ덯삳뭐꣤덑ꭏ쁉ꑈꪺꖭꞡ꙾쓖결랳ꚳ쏶ꅃꕴꕾꅁ녱띾냈ꑈ귻ꖭꞡꩁ냈꙾룪낾ꝃ뭐ꥴ꣠ꭏ돦ꓱ뉶낾낪ꪺ뉻뙈꣓곝ꅁ꣤ꥷ뗛뉶ꕇꚳ뒣ꪺꖲ굮ꅃ ꫭ2₫䒵샅?욪몱풭窲캭? 엜볆ꙗ뫙 ꖭꞡ볆 볐럇깴 돌ꑪ귈 돌ꑰ귈 ꒤ꛬ볆
궷쁉뫞뉺뻇돸 351닄꒭ꣷ 닄ꑔ듁 2003꙾11ꓫ 덑ꭏ쁉ꑈ꙾쓖(랳) 23굮ꭏꑈ꙾쓖(랳)ꕄ곹ꭏ쁉썂(꒸) 500000꙾쎺ꕄ곹ꭏ뙏(꒸) 11354꙾쎺곹ꭏ뙏(꒸) 2160곹ꭏ뙏ꓱ꣒(%) 띾냈ꑈ귻꙾쓖(랳) 32ꩁ냈꙾룪(꙾) 볒ꮬ륷듺꿠ꑏ꒧ꓱ룻 ꖻ곣ꡳꙢ냵ꛦ엞뿨꽓끪쉫껉ꅁ꓀ꝏꕈꅂꅂꅂꝀ결셻곉귈ꅁꣃꯘꗟꕼ귓볒ꮬꅁ꣤꒤ꫭꗜꙢꡓꚳ엧ꭥ뻷뉶ꑕ녠ꗎꪺ볆귈ꅁꛜ꧳ꭨ결룓꒽ꕱ닄ꑇꑑ꒭귓ꓫꪺꖢ껄뉶ꓴ럇ꅁꕴꭨ결냪맘쁉띾87꙾ꯗ닄ꑇꑑ꒭귓ꓫ꒧ꖢ껄뉶ꓴ럇ꅁꭨ결귌ꪺ롧엧볆귈ꅁꙕ볒ꮬ곒ꕈꙖꯡ뇸ꗳꚡ덶ꡂ끪쉫ꩫ뽺뿯엣뗛ꪺ륷듺엜볆ꅃ쏶꧳귋뛇뮼쏾꾫롧뫴룴뭐Ꙩ뱨꣧볆덳떲뫴룴ꓨ궱ꅁ귌ꕈꑇ뫘륷듺엜볆뿯꣺ꓨꩫꅁ꓀ꝏ결롧ꗑ엞뿨꽓끪쉫뽺뿯뭐꿇ꑊꗾ뎡륷듺엜볆ꅆꑇ뫘쇴싃뱨륂뫢꒸볆ꗘ덝ꥷꓨꩫꅁ꓀ꝏ결녎뿩ꑊ뱨뭐뿩ꕘ뱨ꪺ륂뫢꒸볆ꕛ셠ꯡꖭꞡ뭐뙽껚뢹ꑇ뫘ꅁꚹꯘꗟꕼ귓귋뛇뮼쏾꾫롧뫴룴뭐ꕼ귓Ꙩ뱨꣧볆덳떲뫴룴ꅁ돌ꯡꣃꝑꗎ덮엩꧒뒣꣑ꪺꛛ냊뫴룴덝군ꕜ꿠ꯘꗟ볒ꮬꅁꙀ군룕엧ꑑꑔ뫘볒ꮬꅃ곣ꡳ떲ꩇ땯뉻ꅁꝑꗎ쏾꾫롧뫴룴ꪺ쇴싃뱨륂뫢꒸볆ꕛ셠ꯡꖭꞡꓨꩫꯘꗟꪺ뫴룴볒ꮬꅁꗑ꧳쇴싃뱨ꪺ륂뫢꒸룻Ꙩꅁꙝꚹ꣏뫴룴볒ꮬ늣ꗍ끖뵭뭾깴뭐엧쏒뭾깴ꝥ뉻ꕘ깴늧륌ꑪꪺ륌ꯗ뻇닟ꅝover learningꅞ뉻뙈ꅁꫭꗜ룓뫴룴볒ꮬ꣣ꚳ륷듺끖뵭뵤꣒ꪺ꿠ꑏꅁ꿊ꕆ륷듺듺룕볋ꖻꪺ꿠ꑏꅃ걇녎꣤꒤ꕼ귓볒ꮬ뇆낣ꯡꅁꕈꑅ귓볒ꮬ뙩ꛦ듺룕떲ꩇ꒧뭾Ꝑ뉶ꪺꓱ룻ꅃꙕ볒ꮬ꒧걛멣뭐슲뫙꙰ꫭ4꧒ꗜꅁ쏾꾫롧뫴룴냑볆덝ꥷ꒤ꪺ닄ꑀ귓볆꙲ꫭꗜ뿩ꑊ뱨ꪺ륂뫢꒸귓볆ꅂ꣤ꚸ걏쇴싃뱨ꪺ륂뫢꒸귓볆뿩ꕘ뱨ꪺ륂뫢꒸귓볆ꅃ ꫭ5결ꙕ볒ꮬ끖뵭뭐듺룕떲ꩇ꒧귓ꝏ뭐뻣엩ꪺ뭾Ꝑ뉶꒧ꓱ룻ꅁ녱끖뵭뭐듺룕ꪺ떲ꩇꣃꡓꚳꓓꑪ깴늧꣓곝ꅁꙕ볒ꮬ곒ꚳꯜꙮꪺ뒶륍꧊ꅁ엣ꗜꝑꗎ끖뵭ꯡꪺ볒ꮬ꣓듺룕띳ꪺ룪껆ꅁ꣤듺룕껄ꩇꗧꚳ곛Ꙑꪺ륷듺꿠ꑏꅁꗑꚹꕩꪾꖻ곣ꡳ꧒뇄ꗎꪺ륷듺엜볆꣣ꚳꯜꙮꪺ룑쓀꿠ꑏꅃ꣤ꚸꅁꗑ엞뿨꽓끪쉫떲ꩇ녯ꪾꅁꙝ결셻곉귈ꪺ꒣Ꙑꅁ라늣ꗍ꒣Ꙑ땻ꯗꪺꮬI뭐ꮬII뿹뭾ꅁꛓꕂ꣢ꝥ뉻ꓨꙖ엜냊ꪺ쏶ꭙꅃꙝꚹ결궰ꝃ엩뻣ꪺ뭾Ꝑ뉶ꅁ둎ꖲ뚷ꕈ뱗ꕛꮬI뿹뭾Ꝁ결ꕎ믹ꅃ뒫ꕹ룜뮡ꅁ꙰ꩇ귌꣏ꗎ엞뿨꽓끪쉫꣓ꯘꗟ륷듺볒ꮬ껉ꅁ
352맘쁉ꭏ돦ꚭ듁ꖢ껄꒧륷듺 ꒣뫞걏ꕈ룓꒽ꕱꪺ걏맘쁉띾ꪺ꒧엧ꭥꖢ껄뉶ꓴ ꫭ3₵샅?욪몤삧䞱ꆪ? 엜볆ꙗ뫙 떧볆ꓱ꣒(%)엜볆ꙗ뫙 떧볆 ꓱ꣒(%)닄ꑇꑑ꒭귓ꓫ걏ꝟꖢ껄 쎺뙏ꓨꚡ €? ꛛ냊신녢 4407 ₧? 뚰엩띊쎺 438 ꚬ뙏귻ꚬ뙏 7295 덑ꭏ쁉ꑈ꧊ꝏ 쎺뙏꙾듁 ꡫ 6꙾ 910 ꑫ 10꙾ 436 15꙾ 435 20꙾ 10359 덑ꭏ쁉ꑈ녂ꯃ ꚳ때곹 ꑶ녂 ꚳ 8592 ꖼ녂 때 3548 덑ꭏ쁉ꑈ슾띾쏾ꝏ ꚳ때엩샋 닄1쏾 ꚳ 0 0 닄2쏾 때 12140 100 닄3쏾 닄4쏾 닄5쏾 닄6쏾 굮ꭏꑈ꧊ꝏ 걏ꝟ결볐럇엩 ꡫ 걏 12139 100 ꑫ ꝟ 1 0굮ꭏꑈ뭐덑ꭏ쁉ꑈ쏶ꭙ 띾냈ꑈ귻뇐땻ꯗ ꑬꑫ 냪ꑰ 41 ꕓꥮꥦ 냪꒤ 922 끴낸 낪꒤ 7136 ꖻꑈ 녍곬 3078 ꓷꗀ ꑪ뻇 906 깡ꑈ 곣ꡳ꧒ 57 ꣤ꕌ 걏ꝟ결귬ꭏꓡ 띾냈ꑈ귻꧊ꝏ 걏 ꡫ 5062 ꝟ ꑫ 7078 쎺뙏ꝏ 걏ꝟ결ꥴ꣠ꭏ돦 ꙾쎺 걏 9367 ꕢ꙾쎺 ꝟ 2773 ꥵ쎺 ꓫ쎺 럇Ꝁ결셻곉귈ꅁ뎣때ꩫ녎ꮬI뿹뭾궰ꛜ20%ꕈꑕꅃ럭귌ꕈ볒샀ꪺ결셻곉귈껉ꅁ땯뉻꣤떲ꩇ뮷룻ꭥ결꣎ꅁꮬI뿹뭾ꕩ궰ꛜ곹15%ꖪꕫꅁ꙰
궷쁉뫞뉺뻇돸 353닄꒭ꣷ 닄ꑔ듁 2003꙾11ꓫ ꛳ꡍꥷ돌꣎ꪺ셻곉귈ꭨꖲ뚷롧ꗑ놴꿁뭐룕엧꣺녯ꅃꙁ둎귋뛇뮼쏾꾫롧뫴룴ꛓꢥꅁꝑꗎ덮엩ꛛ냊뫴룴덝군ꪺꕜ꿠ꅁ쇶땍믝굮ꫡ뙏룻껉뚡ꅁꕩꕈ쇗ꝋꑈ결ꙝ꿀맯볒ꮬ냑볆덝ꥷꪺ뱶암ꅁꛓꕂꕩꕈ늣ꗍꥍ볒ꮬ둘ꕇꑀ볋ꪺ떲ꩇꅁ꣤ꮬIꅂꮬII뻣엩ꪺ뭾Ꝑ뉶곒Ꙣ20%ꕈꑕꅃ ꫭ4₼튫견宺掻ꆩ? 볒ꮬ뫘쏾 엜볆뿯꣺ꓨꩫ 냑볆덝ꥷ 볒ꮬ슲뫙 LR Ꙗꯡ뇸ꗳꚡ 셻곉귈= LR Ꙗꯡ뇸ꗳꚡ 셻곉귈 Ꙗꯡ뇸ꗳꚡ 셻곉귈 LR Ꙗꯡ뇸ꗳꚡ 셻곉귈 엞뿨꽓끪쉫31-5-1 MFLN-LR MFLN ꗾ뎡엜볆 41-6-1MFLN-allBPN 엞뿨꽓끪쉫22-4-1 BPN-LR BPN ꗾ뎡엜볆 31-5-1BPN-allBPN ꛛ냊뿯꣺20-5-1 BPN-auto ꫭ5₦喼튫겻级傲皤? 돦ꛬ(%) 끖뵭떲ꩇ 듺룕떲ꩇ 볒ꮬ슲뫙 ꮬI ꮬII 뻣엩 ꮬI ꮬII 뻣엩 뿹뭾 뿹뭾 뿹뭾 뿹뭾 뿹뭾 뿹뭾 MFLN-LR MFLN-all BPN-LR BPN-all BPN-auto 귌ꪾ륄ꮬI뿹뭾꧒덹ꚨꪺꚨꖻ띬ꖢ룻결쑙궫ꅁꙝꚹꖻ곣ꡳ껚뻚ꚡ(4)ꅁꓱ룻ꮬI뿹뭾돌ꝃ꒧ꅂMFLN-all뭐BPN-auto떥볒ꮬꅁ샋ꥷ꣤륷듺꿠ꑏ걏ꝟꚳ엣뗛ꪺ깴늧ꅁ꣤떲ꩇ꙰ꫭ6꧒ꗜꅁꙢ5%ꪺ엣뗛ꓴ럇ꑕꅝ|z| >ꅞꅁꝑꗎ엞뿨꽓끪쉫ꅝꅞꓨꩫꝀ륷듺ꅁ꒣뫞걏ꮬIꅂꮬII뻣엩뿹뭾ꅁ꣤떲ꩇ뭐귋뛇뮼쏾꾫롧뫴룴ꅝBPN-autoꅞꣃꡓꚳ엣뗛ꪺ깴늧ꅁꚹꥍꑷꪾꓥ쑭뒶륍뭻결귋뛇뮼쏾꾫롧뫴룴ꓱ엞뿨꽓끪쉫ꚳꟳꙮꪺ륷듺꿠ꑏ꒧떲ꩇ곛ꖪꅁꕩ꿠걏ꙝ셻곉귈덝ꥷ꒣Ꙑ꧒교ꅃꗑꚹꕩꪾꅁ셻곉귈맯꧳볒ꮬ륷듺꿠ꑏꪺ쁵Ꙉ꧊꣣ꚳ쏶쇤꧊ꪺ뱶암ꅃ꣤ꚸꅁ엞뿨꽓끪쉫뭐귋뛇뮼쏾꾫
354맘쁉ꭏ돦ꚭ듁ꖢ껄꒧륷듺 롧뫴룴ꙢꮬI뿹뭾엣뗛ꙡꝃ꧳Ꙩ뱨꣧볆덳떲뫴룴ꅝMFLN-allꅞꅁꙢꮬII뿹뭾ꭨ엣뗛ꙡ룻낪ꅁꙐ껉라뱗ꕛ뻣엩볒ꮬꪺ뭾Ꝑ뉶ꅃꙝꚹꅁ꙰ꩇ돦ꕈ뭾Ꝑ뉶ꪺ낪ꝃ낵결볒ꮬ륷듺꿠ꑏ꒧ꓱ룻ꅁ녎라ꦿ늤ꙝꮬI뿹뭾꧒덹ꚨꪺ띬ꖢꚨꖻꅁ맯맘쁉꒽ꕱꛓꢥꅁ뫫럇ꙡ륷듺ꖢ껄ꭏ돦녎뮷룻뒣ꩀ뻣엩ꪺ륷듺뉶꣓녯궫굮ꅁ걇ꖻ곣ꡳ뗛궫꧳엞뿨꽓끪쉫뭐귋뛇뮼쏾꾫롧뫴룴ꪺꓱ룻ꅃ ꫭ6₼튫겻级傲盀쮩?ꅝz-귈ꅞ † 볒†₫? ꮬI뿹뭾 ꮬII뿹뭾 뻣엩뿹뭾 . BPN-auto . MFLN-all BPN-auto . MFLN-all ꑇ꓀쏾냝썄라ꙝ셻곉귈ꪺ덝ꥷ꒣Ꙑꛓ덹ꚨꖨꑪꪺ깴늧ꅁ걇뙩ꑀꡂꕈAUROC꣓ꓱ룻볒ꮬꪺ냏ꝏ껄ꩇꅁꖻ곣ꡳ땯뉻꧒ꯘꗟꪺ볒ꮬ꒧AUROC곒륆ꅁ엣ꗜꙕ볒ꮬꪺ냏ꝏ껄ꩇꣃ때엣뗛ꪺ깴늧꧊ꅁꚹꕒ꓀뮡ꧺ뭾Ꝑ뉶깴늧ꭙꗑ셻곉귈꒧꒣Ꙑ꧒덹ꚨꅃ뒫귓ꢤꯗ꣓뮡ꅁ굙녎맏1Ꙗ껉쓁신90ꯗꅁꭨꕩ땯뉻꣤결ꮬI뭐ꮬII뿹뭾Ꙣꙕ뫘셻곉귈ꑕ꒧뒲ꝇ맏ꅝ꙰맏2ꅞꅃ 뭾Ꝑꚨꖻ꒧ꓱ룻 둎맘쁉꒽ꕱꛓꢥꅁ뭾녎ꖢ껄ꭏ돦륷듺결ꚳ껄ꅝꮬI뿹뭾ꅞ꧒덹ꚨꪺꮴ곹ꭏꗾꑵꝀ껄뉶궰ꝃ뭐띾냈ꭾ뷨꯹쓲둣꓆꒧ꚨꖻ띬ꖢꅁ라ꓱ뭾녎ꚳ껄ꭏ돦럭ꚨꖢ껄ꅝꮬII뿹뭾ꅞ꣓녯ꑪꅁꙝꛓ꣏녯ꮬI뭐ꮬII뿹뭾ꪺꚨꖻ뉶꒣라ꑰ꧳떥꧳1ꅃ럭ꛒ뱻ꚨꖻ뉶ꪺ엜꓆ꛓ늣ꗍꪺ뭾Ꝑꚨꖻ띬ꖢ껉ꅁ굙돦ꕈ뭾Ꝑ뉶뭐AUROCꪺꓱ룻ꣃ때ꩫꖿ뵔ꙡ뗻꛴볒ꮬꪺꙮ썡ꅃꙝꚹꅁꖻ곣ꡳ뙩ꑀꡂꕈ듺룕볋ꖻꪺ뭾Ꝑꚨꖻ꣓뿅뙱ꙕ볒ꮬꪺ쁵Ꙉ꧊ꅁ뭾Ꝑꚨꖻ띕낪ꅁꫭꗜ볒ꮬ륷듺ꪺ떲ꩇ꧒덹ꚨꪺ띬ꖢ둔ꯗ띕ꑪꅃꙢ꒣Ꙑꪺꚨꖻ뉶ꑕꅁꑅ귓볒ꮬꪺ뭾Ꝑꚨꖻꙃ꧳ꫭ7ꅁꗑꫭ녯ꪾ럭ꚨꖻ뉶ꑪ꧳2껉ꅁꕈ볒ꮬ뭐BPN-autoꪺ륷듺꿠ꑏ룻결뉺띑ꅃ맪ꑗꅁꕴ녱ꫭ6ꗧꕩ땯뉻꣢볒ꮬꪺ뻣엩뭾Ꝑ뉶ꣃꡓꚳ엣뗛ꙡ깴늧ꅝz귈=ꅞꅃ낣ꚹ꒧ꕾꅁ녱ꫭ7귌땯뉻BPN-autoꪺ뭾Ꝑꚨꖻ꒣라쁈뗛ꚨꖻ뉶ꪺ꒣Ꙑꛓꚳꓓꑪꪺ엜ꅝ꣤ꮬIꥍꮬII뿹뭾꓀ꝏ결ꥍꅞꅁꙝ결껚뻚ꚡ(5)ꅁ럭ꮬI뭐ꮬII뿹뭾곛떥껉ꅁ뭾Ꝑꚨꖻ결ꑀ녠볆ꅃꟳ뙩ꑀꡂꛓꢥꅁ굙ꮬI뿹뭾ꑰ꧳ꮬII뿹뭾ꅁꭨ뭾Ꝑꚨꖻ녎쁈ꚨꖻ뉶뱗ꕛꛓ뮼듮ꅆ꒧ꅁꭨ뮼뱗ꅃ
궷쁉뫞뉺뻇돸 355닄꒭ꣷ 닄ꑔ듁 2003꙾11ꓫ ꮬ⁉䤠뿹뭾⠥??㠰㘰㐰㈰??〴〶〸??ꮬ⁉₿僚縨?? 맏2₫?I뭐ꮬII뿹뭾ꚱ뵵맏 ꖻ곣ꡳꕴ샋ꥷ볒ꮬ뭐BPN-autoꙢ꒣Ꙑꪺꚨꖻ뉶ꑕꅁ꣤뭾Ꝑꚨꖻ걏ꝟ꙳Ꙣ엣뗛ꙡ깴늧ꅃꫭ8ꪺꖭꞡ볆뭐볐럇깴ꭙ껚뻚BPN-autoꙢ끖뵭뭐엧쏒뭾깴ꚬ샄뇸ꗳꑕꅁꡃꚸꪺ뻇닟둠샴결20ꅂ뻇닟덴ꯗ결ꅂ멄꧊ꙝꑬ결껉Ꝁ50ꚸꪺ끖뵭ꯡꅁ녎듺룕떲ꩇ꧒늣ꗍꪺꮬI뭐ꮬII뿹뭾녡ꑊꚡ(5)꧒ꡄ녯ꅁꣃ녎꒧뭾Ꝑꚨꖻ덝결뗪때낲덝ꪺꖭꞡ귈ꅁ샋ꥷ떲ꩇ엣ꗜꅁ꣢볒ꮬꪺ뭾Ꝑꚨꖻꣃ꒣라쁈뗛ꚨꖻ뉶ꪺ꒣Ꙑꛓꚳ엣뗛ꙡ깴늧ꅃ 결꿠ꟳ뉍랡ꙡ뮡ꧺ볒ꮬ뭐BPN-autoꪺ쁵Ꙉ꧊ꅁꖻ곣ꡳꕈBPN-auto끖뵭50ꚸꯡ꧒늣ꗍꪺꮬI뭐ꮬII뿹뭾쎸뭳꧳맏3ꅁꣃꕈ뛪쉉ꫭꗜꅁꑔꢤꮬ결꣤볋ꖻꖭꞡ볆ꅆꚱ뵵ꭨ결엞뿨꽓끪쉫꒧ꮬI뭐ꮬII뿹뭾ꅁ녱맏꒤ꕩꕈ땯뉻ꅁ뻣엩ꛓꢥꅁ귋뛇뮼쏾꾫롧뫴룴ꣃꡓꚳ떴맯ꪺ룻꣎륷듺꿠ꑏꅁ맪ꑗꅁꚱ뵵뭐맏쉉꧒ꚨꪺ뗪뵵늣ꗍꗦꑥꅁꫭꗜ꣢뫘볒ꮬꪺ쁵Ꙉ꒬ꢣꅃꪺ셻곉귈ꭙꕈ룕뭾ꩫꛓ녯ꅁꣃ때ꑀꥷ꒽ꚡꕩ녯돌꣎셻곉귈ꅁ걇꣏ꗎꑗ꒴ꕈ귋뛇뮼쏾꾫롧뫴룴룻결슲꧶ꕂ뒶덱ꅃ뫮Ꙙꕈꑗꅁ엽귌맯엞뿨꽓끪쉫뭐쏾꾫롧뫴룴떥꣢ꓨꩫ꒧삳ꗎ뭐ꓱ룻ꚳꑀ룻ꞹ뻣ꕂ뵔맪ꪺ뭻쏑뭐셁룑ꅃ
356맘쁉ꭏ돦ꚭ듁ꖢ껄꒧륷듺 ꫭ7₦喼튫겪못级傦ꢥ? ꚨꖻ뉶 볒ꮬ슲뫙 1 2 3 4 5 10 30 50 MFLN-LR MFLN-all BPN-LR BPN-all BPN-auto ꫭ8₼튫?ꥍBPN-auto꒧뭾Ꝑꚨꖻ샋ꥷ ꚨꖻ뉶 뗪때낲덝 ꖭꞡ볆 볐럇깴 z-귈 1 뭾Ꝑꚨꖻ= 2 뭾Ꝑꚨꖻ= 뭾Ꝑꚨꖻ= 4 뭾Ꝑꚨꖻ= 뭾Ꝑꚨꖻ= 10 뭾Ꝑꚨꖻ= 뭾Ꝑꚨꖻ= 50 뭾Ꝑꚨꖻ=ꮬ⁉?䤠?뿹뿹뭾뭾⠥⠥?⤠㌰㈵㈰ㄵ??ꮬ⁉₿僚縨┩??ㄵ㈰㈵㌰㌵䉐中??瑯䱒ⴰ⸲?浥慮映????慵瑯 맏3?뭐BPN-autoꪺꮬI뭐ꮬII뿹뭾뒲ꝇ맏
궷쁉뫞뉺뻇돸 357닄꒭ꣷ 닄ꑔ듁 2003꙾11ꓫ ꖻ곣ꡳ롧ꗑ뭾Ꝑ뉶ꅂAUROC뭐뭾Ꝑꚨꖻꪺ샋ꥷꯡ땯뉻ꅁ쇶땍귋뛇뮼쏾꾫롧뫴룴ꪺ륷듺꿠ꑏꣃꡓꚳ엣뗛ꙡ쁵꧳엞뿨꽓끪쉫ꅝBoritz and Kennedy, 1995ꅂBoonyanunta and Zeephongsekul, 2000ꅞꅁꗑ맏4ꕩꪾ쏾꾫롧뫴룴뭐엞뿨꽓끪쉫Ꙣꭏ돦ꖢ껄륷듺ꪺ냏ꝏ껄ꩇꑗ꒴ꚳ꣤깴늧꙳Ꙣꅁꑝ둎걏뮡ꅁꝐ쉟볒ꮬꪺ쁵Ꙉꣃ때ꩫ돦꿂ꕈ뭾Ꝑ뉶AUROCꝀꓱ룻ꅃꕴ둎Ꙩ뱨꣧볆덳떲ꮬ뫴룴ꛓꢥꅁ꣤때ꩫꥍꓥ쑭끏룼ꚳ곛Ꙑ껄ꩇꅁꕩ꿠걏ꙝꖻ곣ꡳ꧒뿯꣺ꪺ엜볆꒤ꚳꯜꙨ걏쏾ꝏ룪껆ꅁꙝꚹ때ꩫ롧ꗑ볆꓆뭐맯볆꓆ꪺ덂뉺ꛓ뒣볒ꮬꪺ륷듺꿠ꑏ꧒교ꅃ 뱶암ꭏ돦ꚭ듁ꖢ껄꒧엜볆뮡ꧺ 껚뻚꧒곣ꡳ꒽ꕱ꒧ꗾ꙾ꯗꭏ돦ꅂꭏꓡꅂ껖ꭏ뭐띾냈ꑈ귻떥ꙝ꿀ꯘꗟꪺ엞뿨꽓끪쉫꛴군떲ꩇ꙰ꫭ9꧒ꗜꅃ궺ꗽꅁꙢꭏꓡꙝ꿀ꓨ궱ꅁ덑ꭏ쁉ꑈꪺ꙾쓖띕낪ꅁꭏ돦ꖢ껄ꪺ뻷뉶띕ꝃꅆꛓ덑ꭏ쁉ꑈ결굮ꭏꑈ꒧ꑰꯄꅂꓷꗀꅁꭨ꣤ꖢ껄ꪺ뻷뉶ꗧ룻ꝃꅁꙝꙢ덯뫘쏶ꭙꑕ꧒쇊뙒ꪺꭏ돦꒺깥덱녠ꕈꭏ믙ꕜ꿠결ꕄꅁ걇룻꒣라뮴꧶ꙡ룑곹ꅆꕴ꒽ꕱ귬ꭏꓡ꧒뱗쇊ꪺꭏ돦ꗧꓱ룻꒣라룑곹ꅁꚹꕒ꓀ꙡ엣ꗜ꣺녯ꭏꓡ맯꒽ꕱ냓ꭾ뭐ꩁ냈ꭾ뷨ꪺ뭻Ꙑꅁꑾ라꣏꣤쑾쓲쇊뙒꒽ꕱꪺ냓ꭾꅃ ꣤ꚸꅁꙢꭏ돦꽓꧊ꓨ궱ꅁꭏ돦뇄꙾쎺ꕢ꙾쎺쎺뙏ꓨꚡꅁꕈꝑꗎ믈ꛦ신녢뚰쎺ꓨꚡ쎺ꗦꭏ뙏ꅁꕂ쎺뙏꙾듁띕땵ꅁ꣤ꖢ껄ꪺ뻷뉶룻ꝃꅁꑀ꿫꣓뮡ꅁ쎺뙏ꚸ볆띕ꓖ띕꒣꧶땯ꗍꟑ끏쎺뙏ꪺ냝썄ꅁꛜ꧳ꛛ냊신녢뭐뚰쎺ꓨꚡꗧ꿠ꙝꚳꭏ뙏ꚩꛓꝬꓞꭏꓡꪺꩠ띎ꅁ뙩ꛓ궰ꝃꭏ돦ꖢ껄ꪺ뻷뉶ꅆ쎺뙏꙾듁띕땵ꅁꭨꭏꓡ굴뻡ꭏ뙏듁뚡룻땵ꅁꙝꚹ덱녠꒣쑀ꙝ룑곹ꛓ덹ꚨ롧샙ꑗꪺ띬ꖢꅆꙢꕴꑀꓨ궱ꅁꭏ썂ꅂ꙾쎺ꭏ뙏띕낪ꕈꚳ곹ꪺꭏ돦ꅁꭏꓡꕩ꿠ꙝ결ꚬꑊꪺ엜꓆ꛓ때ꩫ굴뻡ꅁ꣏녯ꭏ돦룑곹ꖢ껄ꪺ뻷뉶룻낪ꅆꛓ곹ꪺꭏ뙏룻낪ꅁ덱녠ꚹ쏾ꭏ돦ꚳ룻낪ꪺꗍꭥ떹ꕉꭏ믙ꅁꙝꚹ룻꒣라뮴꧶ꙡ룑곹ꅃ 띾냈ꑈ귻ꙝ꿀ꓨ궱ꅁꑫ꧊꙾룪룻낪ꪺ띾냈ꑈ귻ꅁ꣤ꭏ돦ꖢ껄ꕩ꿠꧊룻낪ꅁ덯ꕩ꿠걏ꙝ결ꑫ꧊띾냈ꑈ귻꒤ꅁ꣓ꛛ꧳뗦쑸뇚ꪺꑈ놡ꭏꛓꭄꕈ뙱ꢭ덝군ꪺꭏ쁉덗릺꧒교ꅆꛓ꙾룪룻낪ꪺ띾냈ꑈ귻ꕩ꿠걏ꙝ결ꩁ냈ꪺ멁ꯗ뭐볶ꟕ꒣릳꙾뮴띾냈ꑈ귻꿫ꪺ뽮랥뭐볶놡ꅁꕂꕩ꿠ꙝ녍띾ꪾ쏑ꪺꝬꚬ룻멃ꛓ때ꩫ몡ꢬ껸뙏Ꙩ꒸꓆ꪺ믝ꡄ꧒교ꅃ돌귈녯띾ꩠ띎ꪺ걏뱶암ꭏ돦ꚭ듁ꖢ껄ꪺ쏶쇤ꙝ꿀결띾냈ꑈ귻싷슾룵병ꅁꚹ녎덹ꚨꭏ돦ꖢ껄ꪺꕩ꿠꧊ꑪꑪꙡ뒣ꅁ꣤ꕄꙝꭙ띾냈ꑈ귻룵병싷슾ꛓ늣ꗍꭏꓡ맯ꩁ냈ꭾ뷨ꪺ꒣몡띎ꅁꙝꛓ뒣낪룑곹ꪺ뻷뉶ꅃ ꛜ꧳꣏ꗎ쏾꾫롧뫴룴ꓨ궱ꅁꖻ곣ꡳꕈ뇓띐ꯗ꓀꩒ꅝsensitivity analysisꅞ꣓뮡ꧺ뿩ꑊ엜볆꒧곛맯궫굮꧊ꅃ뇓띐ꯗ꓀꩒ꭙ럭Ꝓ낣걙ꑀ뿩ꑊ엜볆껉꧒뻉교볒ꮬ듺룕떲ꩇ뱗ꕛ꒧뭾깴ꛊ꓀ꓱꅁꚹ뮡ꧺ룓뿩ꑊ엜볆ꪺ곛맯궫굮꧊ꅁ
358맘쁉ꭏ돦ꚭ듁ꖢ껄꒧륷듺 뭾깴ꓱ뉶띕ꑪꫭꗜ룓뿩ꑊ엜볆ꪺ뱶암ꑏ띕ꑪꅃ둎쏾꾫롧뫴룴ꪺ떲ꩇꛓꢥꅁꗑ꧳덮엩Ꙣꯘꗟ볒ꮬ껉ꅁ굙결ꙗꗘ엜볆ꭨ라ꛛ냊ꕈ뗪샀엜볆ꕛꕈ덂뉺ꅁꙝꚹꙢ뿩ꑊ엜볆껉ꅁ때뚷녎룓엜볆신뒫ꚨ뗪샀엜볆ꅃꗑꚹꕩꪾꅁꙢ뙩ꛦ뇓띐꧊꓀꩒껉ꅁ꣤떲ꩇ꒴ꕈꙗꗘ엜볆결뿩ꕘ떲ꩇꅁ뒫ꕹ룜뮡ꅁ덑ꭏ쁉ꑈ뭐굮ꭏꑈ꒧쏶ꭙꅂ쎺뙏ꝏ뭐쎺뙏ꓨꚡ떥륷듺엜볆ꅁ귌꓀ꝏꕈx-xꅂx-xꅂx-x71214161718꣓ꫭꗜ덯ꑔ귓엜볆ꅃ ꫭ9₼튫?꓀꩒떲ꩇ 룑쓀엜볆 ꭙ볆꛴군 볐럇뭾 Waldꕤꓨ귈 p-귈 ** x * *** * *** x *** *** x *** *** x *** *** x *** *** x *** *** x *** *** x *** *** 녠볆뚵 -2 Log Likelihood 2Cox & Snell – R .513 2Nagelkerke – R .704 *ꅁ*****꓀ꝏꫭꗜ엣뗛ꓴ럇α=ꅁꅃ 껚뻚ꫭ10ꕩꕈ땯뉻ꅁ엜볆궫굮꧊ꪺ땻ꯗ뇆Ꟈꅁꕈ띾냈ꑈ귻ꪺ싷슾룵병맯ꭏ돦걏ꝟꖢ껄ꪺ뱶암돌결엣뗛ꅁ룓엜볆ꪺ뭾깴ꓱ뉶ꫭꗜ럭볒ꮬ뇆낣엜볆x껉ꅁ녎뻉교볒ꮬ듺룕떲ꩇ꒧뭾깴귈뱗ꕛ%ꅃ꣤ꚸ걏ꭏ돦ꪺ쎺31뙏ꓨꚡꅂꚳ때곹ꅂ띾냈ꑈ귻꙾쓖뭐ꑵꝀ꙾룪떥ꅁꕈꑗꥍ엞뿨꽓끪쉫ꪺ떲
궷쁉뫞뉺뻇돸 359닄꒭ꣷ 닄ꑔ듁 2003꙾11ꓫ ꩇꚳ낪ꯗꪺꑀ교꧊ꅃꚹꕩ뇀ꪾꅁ꙾듁맘쁉ꭏ돦ꚭ듁ꖢ껄ꪺꙝ꿀꒤ꅁꕈ띾냈ꑈ귻걏ꝟꙢ슾ꅂ꒽ꕱꪺꩁ냈ꭾ뷨뭐ꭏ돦냓ꭾꪺ덝군떥ꪺ뱶암돌결엣뗛ꅁꕴꭏ쁉썂ꅂꭏ뙏뭐굮ꭏꑈ덑ꭏ쁉ꑈ꒧롧샙ꪬꩰꗧꚳꯜꑪꪺ곛쏶꧊ꅃ귌ꖲ뚷뮡ꧺꪺ걏ꅁ뇓띐꧊꓀꩒ꕵ꿠셁룑엜볆곛맯ꪺ궫굮꧊ꅁꛜ꧳꣤ꙝꩇ쏶ꭙꭨ때ꩫ녯ꪾꅃ ꫭ10ﺯꮸ枺?ꅝBPN-autoꅞ꒧뇓띐꧊꓀꩒ 엜볆 x-x x-xx-xxxxxxxx x x 7121416 1718 192021222428 293031ꓱ뉶 뇆Ꟈ 9 8 2 7 10 12 3 6 4 5 11 1 5. 떲뷗뭐ꯘ쒳 맘쁉꒽ꕱꪺ꙾듁ꭏ돦굙ꚭ듁ꖢ껄맯뻣엩롧샧ꪺ뱶암걊뉠ꕂ뮷ꅁ걇맯ꚭ듁ꖢ껄ꪺ륷듺맪결ꕄ뫞럭ꞽ곆ꛜ띾ꖻꢭ꧒삳궫뗸ꪺ뷒썄ꅃ껚뻚ꖻ곣ꡳ꧒뇄ꗎꪺ볒ꮬꓱ룻녯ꪾꅁ엞뿨꽓끪쉫ꅝꅞꅂ귋뛇뮼쏾꾫롧뫴룴ꅝBPN-autoꅞ뭐Ꙩ뱨꣧볆덳떲뫴룴ꅝMFLN-allꅞ떥ꑔ귓볒ꮬꙢ뻣엩ꪺ륷듺꿠ꑏ뎣ꭄ녠꣣ꚳ뮡ꩁꑏꅁ꣤ꖿ뵔뉶곒륆ꛊ꓀꒧ꑋꑑꑇꕈꑗꅁꕂ볒ꮬꪺAUROC귈낪륆ꅁ엣ꗜ볒ꮬꪺ냏ꝏ껄ꩇ곆꣎ꅁꥍ륌ꕨ곣ꡳ뒶륍뭻결쏾꾫롧뫴룴ꚳ룻꣎ꪺ륷듺껄ꩇꚳ꧒깴늧ꅁꕄ굮ꪺ귬ꙝꙢ꧳셻곉귈ꪺ덝ꥷꛓ늣ꗍꪺ깴늧꧒교ꅃ걇귌ꯘ쒳뇄ꗎAUROC꣓ꓱ룻ꑇ꓀쏾냝썄ꪺ륷듺껄ꩇꅁꙝ결꣤꒣꣼셻곉귈ꪺ뱶암ꅃ걏ꓱ룻AUROCꪺꑪꑰ때ꩫꝊꩁ볒ꮬꙢ맪믚륂ꗎ껉궱셻ꚨꖻ뉶엜꓆ꪺ냝썄ꅁꙝꚹꖻ곣ꡳ뙩ꑀꡂꝑꗎ뭾Ꝑꚨꖻ꣓뮡ꧺ볒ꮬꪺ쁵Ꙉ꧊ꅃꗑ맪뗽ꪺ떲ꩇ녯ꪾꅁ꓀쏾륷듺볒ꮬꙢ꒣Ꙑꪺ셻곉귈뭐ꚨꖻ뉶ꑕꅁ라늣ꗍ꒣Ꙑꪺ뭾Ꝑꚨꖻꅁ뉺뷗ꑗꕩꕈꙢ꽓ꥷꚨꖻ뉶ꑕꡄ녯뭾Ꝑꚨꖻ결돌ꝃꪺ셻곉귈ꅁꚨꖻ뉶ꪺꑪꑰꯜ쏸뫫뵔ꙡ덑꛴뫢ꅁꙝꚹꖻ곣ꡳ볒샀Ꙣ꒣Ꙑꪺꚨꖻ뉶ꑕꅁꡄ녯뭾Ꝑꚨꖻ룻ꑰꪺ륷듺볒ꮬꅁꕈꚳ껄ꙡ궰ꝃ볒ꮬꙝ샴맒엜꓆ꛓ늣ꗍꪺ뭾Ꝑ띬ꖢꅃ 귌ꪾ륄돦ꕈꖿ뵔륷듺뉶ꪺ낪ꝃ꣓뿅뙱볒ꮬꪺ쁵Ꙉ녎ꚳ압ꚹꖢꦼꪺ놡ꩰ땯ꗍꅁ둎꙰ꓥ꒤꧒굺ꅁꗴ꛳볒ꮬꪺ륷듺뎣녎늣ꗍꮬI뭐ꮬII뿹뭾ꅁꯪꖩꚹ꣢볐ꚨ결ꓱꪺ껄삳ꅁ덯ꖿ걏ꑪꙨꓥ쑭ꥼꖼ뉠ꑊ놴끑귈녯궫뗸꒧덂ꅃꗑꖻ곣ꡳ녯ꪾ쏾꾫롧뫴룴ꪺ냏ꝏ껄ꩇꣃꡓꚳ엣뗛ꙡ쁵꧳엞뿨꽓끪쉫ꅁ엞뿨꽓끪쉫ꕈ엧ꭥ뻷뉶낵결셻곉귈ꪺ낵ꩫ깥꧶늣ꗍ룻ꑪꪺꮬI뿹뭾ꅁꙁꅁ뛇닎닎군ꓨꩫꪺ꣏ꗎꗧ꙳Ꙣ뗛덜Ꙩꪺ궭꣮ꅃꛜ꧳쏾꾫롧뫴룴ꭨꕩꝊꩁ룪껆꓀끴ꖲ뚷닅Ꙙ꽓ꥷ꓀끴ꪺ냝썄ꅁꛓꕂ빁ꗎ꧳ꑪ뙱룪껆ꪺ륂뫢ꅃ쇶땍
360맘쁉ꭏ돦ꚭ듁ꖢ껄꒧륷듺 쏾꾫롧뫴룴ꑝꚳ뫴룴걛멣ꅂ뻇닟덴ꯗꅂ쇴싃뱨돦꒸볆엜볆뿯꣺떥뷒썄뚷Ꝋꩁꅁ덺륌뻇닟뭐롧엧닖뽮ꅁ쏾꾫롧뫴룴ꛛ꒣ꖢ결ꑀ뫘덂뉺ꭄ뵵꧊쏶ꭙꪺ돌꣎뿯뻜ꅃ Ꙣꓱ룻귋뛇뮼쏾꾫롧뫴룴ꅝBPN-autoꅞ뭐엞꽓꽓끪쉫ꅝꅞꪺ엣뗛륷듺엜볆ꯡ땯뉻ꅁ꣢ꪺ떲ꩇ꣣ꚳ곛럭낪ꪺꑀ교꧊ꅁ낣ꑆꫭꗜ꧒뿯ꗎꪺ륷듺엜볆꣣ꚳꯜꙮꪺ룑쓀꿠ꑏꕾꅁꣃꕩꝑꗎ엞뿨꽓끪쉫볒ꮬꪺ엣뗛엜볆꣓Ꝋꩁ쏾꾫롧뫴룴때ꩫ뮡ꧺ엜볆뚡ꪺꙝꩇ쏶ꭙ꒧냝썄ꅃ껚뻚맪뗽떲ꩇ땯뉻ꅁ뱶암맘쁉ꭏ돦ꚭ듁ꖢ껄ꪺꙝ꿀ꕈꭏꓡꅂꭏ돦띾냈ꑈ귻ꑔꓨ궱결ꕄꅁ꙰덑ꭏ쁉ꑈ띕꙾뮴띕ꚳ룑곹ꪺ뛉Ꙗꅆꭄ귬ꚳꪺꭏꓡꗧ룻ꕩ꿠덹ꚨꭏ돦ꖢ껄ꅆ뇄ꓫ쎺ꓨꚡꗧ라뱗ꕛ룑곹ꪺꕩ꿠꧊ꅆ돌궫굮ꪺ걏꒽ꕱꙕ뎡ꪺꩁ냈ꭾ뷨ꅁꓗ꣤걏귬띾냈ꑈ귻꒧룵병싷슾ꅁ맯ꭏ돦ꚭ듁ꖢ껄ꪺ뱶암돌결엣뗛ꅃꙝꚹꅁ맘쁉꒽ꕱ꙰꛳뫻꯹띾냈ꑈ귻ꪺ쎭ꥷ꧊ꅁꣃ뒣ꩀꙕ뚵ꩁ냈ꭾ뷨떥꣢ꓨ궱녎걏룓꒽ꕱꗃ쓲롧샧ꪺ꒣ꑇꩫꅃ 냑ꛒꓥ쑭 1. Ꝧ맅곕ꅁꕸ왗늣쁉늣띾ꭏ쁉끝냈곣ꡳꇐ엞뿨꽓끪쉫ꪺ륂ꗎꅁ냪ꗟ낪뚯닄ꑀ곬ꑪ뻇ꭏ쁉샧륂곣ꡳ꧒뫓ꑨ뷗ꓥꅁ2000꙾ꅃ 2. ꩌ쒣ꩆꅁꑋꑑꑔ꙾ꯗ귓ꑈ맘쁉ꭏ돦닄ꑑꑔ귓ꓫ룑곹ꖢ껄꓀꩒ꅁ맘쁉ꥵꕚꅁ닄109듁ꅁ1998꙾ꅁ궶10-26ꅃ 3. ꩌ쒣ꩆꅁ엳엩꣮떽군땥뭐ꭏ돦쑾쓲뉶ꅁꭏ쁉녍ꕚꅁ닄52뿨ꅁ1998꙾ꅁ궶162-181ꅃ 4. 걟ꭔꅁ맘쁉띾뉍쁶꿠ꑏ륷쒵볒ꮬꇐ쏾꾫롧뫴룴꒧삳ꗎꅁ냪ꗟꕸ왗ꑪ뻇끝냈뿄곣ꡳ꧒뫓ꑨ뷗ꓥꅁ1994꙾ꅃ 5. 남꒤얻ꅁꕸ왗ꙡ냏맘쁉띾뉍쁶꿠ꑏ륷쒵볒ꮬLogit뭐쏾꾫롧뫴룴꒧삳ꗎꅁ덻ꗒꑪ뻇ꭏ쁉뻇곣ꡳ꧒뫓ꑨ뷗ꓥꅁ1996꙾ꅃ 6. 뎯쁁ꟸꅂ덜덱ꙷꅂꩌ붯꾳ꅁ믈ꛦ뇂ꭈꯈꓡ륈곹궷쁉꒧륷듺ꅁ뫞뉺곬뻇뻇돸ꅁ닄13ꣷꅁ닄2듁ꅁ 1996꙾ꅁ궶173-195ꅃ 7. 둞ꓥꖿ쒶ꅁ룪껆뇄쑱ꇐ압ꯈ쏶ꭙ뫞뉺멛륱ꑬꛦ빐꒧삳ꗎꅁꕸꕟꅁ볆돕뫴룪끔ꗷꚳ궭꒽ꕱꅁ2001꙾ꅃ 8. 뢭꧉ꚨꅁ삳ꗎ쏾꾫롧뫴룴ꅁ뺧ꩌꕘꪩꫀꅁ1997꙾ꅃ 9. 램ꥶꅁꕸ왗ꙡ냏귓ꑈ맘쁉ꭏ돦ꚭ듁ꖢ껄꒧곣ꡳꅁ걆ꩶꑪ뻇ꭏ쁉곣ꡳ꧒뫓ꑨ뷗ꓥꅁ1992꙾ꅃ 10. Adams, N. M. and D. J. Hand, 2000, An improved measure for comparing diagnostic tests, Computers in Biology and Medicine, 30, 89-96. 11. Boonyanunta, N. and P. Zeephongsekul, 2000, State of the Art Credit Risk Analysis Model: Comparative Analysis Between Statistical Approaches
궷쁉뫞뉺뻇돸 361닄꒭ꣷ 닄ꑔ듁 2003꙾11ꓫ and Neural Network Approaches, The Fifth Conference of the Association of Asian-Pacific Operations Research Societies within IFORS, Singapore. 12. Boritz, J. E. and D. B. Kennedy, 1995, Effectiveness of Neural Network Types for Prediction of Business Failure, Expert Systems With Application, 9, 4, 503-512. 13. Brockett, P. L., W. W. Cooper, L. L. Golden, and U. Pitaktong, 1994, A Neural Network Method for obtaining an early warning of insurer insolvency, The Journal of Risk and Insurance, 61, 3, 402-424. 14. Carter, P., 1995, United Kingdom-Lapsation and the Policyholder:Buyer Study, Life Insurance Marketing and Research Association. 15. Cox, S. H., P. D. Laporte, S. R. Linney, and L. Lombardi, 1992, Single-Premium Deferred Annuity Persistency Study, Transactions of Society of Actuaries Reports, 281-332. 16. Fausett, L., 1994, Fundamentals of neural networks - architectures, algorithms, and applications, New Jersey: Prentice Hall. 17. Hosmer, D. W. and S. Lemeshow, 1989, Applied Logistic Regression, New York: John Wiley. 18. Haykin, S., 1999, Neural networks: A comprehensive foundation, New Jersey: Prentice Hall. 19. Kelly, T. H., 1996, Long-Term Ordinary Lapse Survey, Life Insurance Marketing and Research Association. 20. Lin, S. L., 1996, Financial Distress Classification in the Life Insurance Industry, Journal of Insurance Regulation, 14, 3, 314-342. 21. Norman, F. B., 1960, First Year Lapse and Default Rate, Transactions, society of Actuaries, 12, 258-293. 22. Outreville, J. F., 1990, Whole-life Insurance Lapse Rates and the Emergency Fund Hypothesis, Insurance, Mathematics & Economics, 9, 249-255. 23. Purushotham, M., 2001, . Individual Life Persistency Study, Life Insurance Marketing and Research Association. 24. Sondergeld, E. T., 1997, Persistency and replacement: A best practices for Taiwan, the Philippines, and Hong Kong, Life Insurance Marketing and Research Association. 25. Tsai, C., W. Kuo, and W. Chen, 2002, Early Surrender and the Distribution of Policy Reserves, Insurance: Mathematics and Economics, 31, 429-445. 26. Zweig, M. H. and G. Campbell, 1993, Receiver-Operating Characteristic (ROC) Plots: A Fundamental Evaluation Tool in Clinical Medicine, Clin Chem, 39, 561-577.
