Olas\u0131l\u0131k oranlar\u0131 en yayg\u0131n olarak, bir veya daha fazla \u00f6ng\u00f6r\u00fcc\u00fc de\u011fi\u015fkene ve ikili yan\u0131t de\u011fi\u015fkenine sahip bir regresyon modeline uymak i\u00e7in kulland\u0131\u011f\u0131m\u0131z bir y\u00f6ntem olan lojistik regresyonda<\/a> g\u00f6r\u00fcn\u00fcr.<\/span><\/p>\n D\u00fczeltilmi\u015f olas\u0131l\u0131k oran\u0131,<\/strong> bir modeldeki di\u011fer \u00f6ng\u00f6r\u00fcc\u00fc de\u011fi\u015fkenler i\u00e7in ayarlanm\u0131\u015f bir olas\u0131l\u0131k oran\u0131d\u0131r.<\/span><\/p>\n Di\u011fer \u00f6ng\u00f6r\u00fcc\u00fc de\u011fi\u015fkenlerin etkisine g\u00f6re ayarlama yap\u0131ld\u0131ktan sonra<\/em> , bir yorday\u0131c\u0131 de\u011fi\u015fkenin bir olay\u0131n meydana gelme \u015fans\u0131n\u0131 nas\u0131l etkiledi\u011fini anlamam\u0131za yard\u0131mc\u0131 olmas\u0131 a\u00e7\u0131s\u0131ndan \u00f6zellikle faydal\u0131d\u0131r.<\/span><\/p>\n A\u015fa\u011f\u0131daki \u00f6rnek, olas\u0131l\u0131k oran\u0131 ile d\u00fczeltilmi\u015f olas\u0131l\u0131k oran\u0131 aras\u0131ndaki fark\u0131 g\u00f6stermektedir.<\/span><\/p>\n Bir annenin ya\u015f\u0131n\u0131n d\u00fc\u015f\u00fck do\u011fum a\u011f\u0131rl\u0131kl\u0131 bir bebe\u011fe sahip olma olas\u0131l\u0131\u011f\u0131n\u0131 etkileyip etkilemedi\u011fini anlamak istedi\u011fimizi varsayal\u0131m.<\/span><\/p>\n Bunu ara\u015ft\u0131rmak i\u00e7in, \u00f6ng\u00f6r\u00fcc\u00fc de\u011fi\u015fken olarak ya\u015f\u0131 ve yan\u0131t de\u011fi\u015fkeni<\/a> olarak d\u00fc\u015f\u00fck do\u011fum a\u011f\u0131rl\u0131\u011f\u0131n\u0131 (evet veya hay\u0131r) kullanarak lojistik regresyon ger\u00e7ekle\u015ftirebiliriz.<\/span><\/p>\n 300 anne hakk\u0131nda veri toplad\u0131\u011f\u0131m\u0131z\u0131 ve lojistik regresyon modeline uydu\u011fumuzu varsayal\u0131m. Sonu\u00e7lar burada:<\/span> <\/p>\n Ya\u015fa g\u00f6re olas\u0131l\u0131k oran\u0131n\u0131 elde etmek i\u00e7in katsay\u0131 tahminini tablodan \u00fcstel hale getirin: e 0,173<\/sup> = 1,189<\/strong> .<\/span><\/p>\n Bu bize ya\u015ftaki bir y\u0131ll\u0131k art\u0131\u015f\u0131n, bebe\u011fin d\u00fc\u015f\u00fck do\u011fum a\u011f\u0131rl\u0131\u011f\u0131na sahip olma ihtimalinde 1.189<\/strong> art\u0131\u015fla ili\u015fkili oldu\u011funu s\u00f6yl\u00fcyor. Ba\u015fka bir deyi\u015fle, ya\u015ftaki her bir y\u0131ll\u0131k art\u0131\u015fla birlikte d\u00fc\u015f\u00fck do\u011fum a\u011f\u0131rl\u0131kl\u0131 bebek sahibi olma \u015fans\u0131 %18,9<\/strong> oran\u0131nda artmaktad\u0131r.<\/span><\/p>\n Bu odds oran\u0131na “kaba” odds oran\u0131 veya “d\u00fczeltilmemi\u015f” odds oran\u0131 denir \u00e7\u00fcnk\u00fc modeldeki tek<\/em> \u00f6ng\u00f6r\u00fcc\u00fc de\u011fi\u015fken oldu\u011fundan modeldeki di\u011fer yorday\u0131c\u0131 de\u011fi\u015fkenler i\u00e7in ayarlanmam\u0131\u015ft\u0131r.<\/span><\/p>\n Ancak bir annenin ya\u015f\u0131n\u0131n ve<\/em> sigara i\u00e7me al\u0131\u015fkanl\u0131\u011f\u0131n\u0131n d\u00fc\u015f\u00fck do\u011fum a\u011f\u0131rl\u0131kl\u0131 bebek sahibi olma olas\u0131l\u0131\u011f\u0131n\u0131 etkileyip etkilemedi\u011fini anlamak istedi\u011fimizi varsayal\u0131m.<\/span><\/p>\n Bunu ara\u015ft\u0131rmak i\u00e7in, \u00f6ng\u00f6r\u00fcc\u00fc de\u011fi\u015fkenler olarak ya\u015f ve sigara i\u00e7meyi (evet veya hay\u0131r) ve yan\u0131t de\u011fi\u015fkeni<\/a> olarak d\u00fc\u015f\u00fck do\u011fum a\u011f\u0131rl\u0131\u011f\u0131n\u0131 kullanarak lojistik regresyon ger\u00e7ekle\u015ftirebiliriz.<\/span><\/p>\n 300 anne hakk\u0131nda veri toplad\u0131\u011f\u0131m\u0131z\u0131 ve lojistik regresyon modeline uydu\u011fumuzu varsayal\u0131m. Sonu\u00e7lar burada:<\/span> <\/p>\n Sonu\u00e7lar\u0131 nas\u0131l yorumlayaca\u011f\u0131n\u0131z a\u015fa\u011f\u0131da a\u00e7\u0131klanm\u0131\u015ft\u0131r:<\/span><\/p>\n Ya\u015f:<\/strong> Ya\u015fa g\u00f6re d\u00fczeltilmi\u015f odds oran\u0131 \u015fu \u015fekilde hesaplan\u0131r: e 0,045<\/sup> = 1,046<\/strong> . Bu, sigara i\u00e7me<\/em> de\u011fi\u015fkeninin sabit kald\u0131\u011f\u0131 varsay\u0131ld\u0131\u011f\u0131nda, ya\u015ftaki her ilave y\u0131ll\u0131k art\u0131\u015f i\u00e7in d\u00fc\u015f\u00fck do\u011fum a\u011f\u0131rl\u0131kl\u0131 bir bebe\u011fe sahip olma \u015fans\u0131n\u0131n %4,6 artt\u0131\u011f\u0131 anlam\u0131na gelir.<\/span><\/p>\n \u00d6rne\u011fin, Anne A ve Anne B’nin her ikisinin de sigara i\u00e7ti\u011fini varsayal\u0131m. E\u011fer anne A, anne B’den bir ya\u015f b\u00fcy\u00fckse, anne A’n\u0131n d\u00fc\u015f\u00fck do\u011fum a\u011f\u0131rl\u0131kl\u0131 bebek sahibi olma olas\u0131l\u0131\u011f\u0131, anne B’nin d\u00fc\u015f\u00fck do\u011fum a\u011f\u0131rl\u0131kl\u0131 bebek sahibi olma olas\u0131l\u0131\u011f\u0131n\u0131n 1.046 kat\u0131d\u0131r.<\/span><\/p>\n Sigara i\u00e7me<\/strong> : Sigara i\u00e7meye ili\u015fkin d\u00fczeltilmi\u015f olas\u0131l\u0131k oran\u0131 \u015fu \u015fekilde hesaplan\u0131r: e.485<\/sup> = 1.624<\/strong> . Bu, ya\u015f<\/i> de\u011fi\u015fkeninin sabit kald\u0131\u011f\u0131 varsay\u0131ld\u0131\u011f\u0131nda, anne sigara i\u00e7iyorsa (sigara i\u00e7memeye k\u0131yasla) d\u00fc\u015f\u00fck do\u011fum a\u011f\u0131rl\u0131kl\u0131 bebek sahibi olma \u015fans\u0131n\u0131n %62,4 artt\u0131\u011f\u0131 anlam\u0131na gelir.<\/span><\/p>\n \u00d6rne\u011fin Anne A ve Anne B’nin her ikisinin de 30 ya\u015f\u0131nda oldu\u011funu varsayal\u0131m. Anne A hamilelikte sigara i\u00e7iyor ve anne B sigara i\u00e7miyorsa, anne A’n\u0131n d\u00fc\u015f\u00fck do\u011fum a\u011f\u0131rl\u0131kl\u0131 bebek sahibi olma \u015fans\u0131, anne B’nin d\u00fc\u015f\u00fck do\u011fum a\u011f\u0131rl\u0131kl\u0131 bebek sahibi olma ihtimalinden %62,4 daha y\u00fcksektir.<\/span><\/p>\n Ya\u015fa g\u00f6re d\u00fczeltilmi\u015f olas\u0131l\u0131k oran\u0131n\u0131n, \u00f6nceki \u00f6rnekteki d\u00fczeltilmemi\u015f olas\u0131l\u0131k oran\u0131ndan daha d\u00fc\u015f\u00fck oldu\u011funu unutmay\u0131n. Bunun nedeni, di\u011fer yorday\u0131c\u0131 de\u011fi\u015fkenler yan\u0131t de\u011fi\u015fkeninin ortaya \u00e7\u0131kma \u015fans\u0131n\u0131 artt\u0131rd\u0131\u011f\u0131nda, modelde halihaz\u0131rda mevcut olan yorday\u0131c\u0131 de\u011fi\u015fken i\u00e7in d\u00fczeltilmi\u015f olas\u0131l\u0131k oran\u0131n\u0131n her zaman azalacak olmas\u0131d\u0131r.<\/span><\/p>\n Olas\u0131l\u0131k oran\u0131<\/strong> (bazen “kaba” olas\u0131l\u0131k oran\u0131 olarak da adland\u0131r\u0131l\u0131r), yorday\u0131c\u0131 de\u011fi\u015fkendeki de\u011fi\u015fikliklerin bir yan\u0131t de\u011fi\u015fkeninin ortaya \u00e7\u0131kma \u015fans\u0131n\u0131 nas\u0131l etkiledi\u011fini bize anlatmada faydal\u0131d\u0131r.<\/span><\/p>\n D\u00fczeltilmi\u015f olas\u0131l\u0131k oran\u0131,<\/strong> bir modeldeki di\u011fer \u00f6ng\u00f6r\u00fcc\u00fc de\u011fi\u015fkenler kontrol edildikten sonra<\/em> , bir yorday\u0131c\u0131 de\u011fi\u015fkendeki de\u011fi\u015fikliklerin, bir yan\u0131t de\u011fi\u015fkeninin ortaya \u00e7\u0131kma \u015fans\u0131n\u0131 nas\u0131l etkiledi\u011fini bize anlatmak a\u00e7\u0131s\u0131ndan faydal\u0131d\u0131r.<\/span><\/p>\n Lojistik Regresyona Giri\u015f<\/a> \u0130statistiklerde olas\u0131l\u0131k oran\u0131 bize bir olay\u0131n tedavi grubunda meydana gelme olas\u0131l\u0131\u011f\u0131 ile bir olay\u0131n kontrol grubunda meydana gelme olas\u0131l\u0131\u011f\u0131 aras\u0131ndaki oran\u0131 anlat\u0131r. Olas\u0131l\u0131k oranlar\u0131 en yayg\u0131n olarak, bir veya daha fazla \u00f6ng\u00f6r\u00fcc\u00fc de\u011fi\u015fkene ve ikili yan\u0131t de\u011fi\u015fkenine sahip bir regresyon modeline uymak i\u00e7in kulland\u0131\u011f\u0131m\u0131z bir y\u00f6ntem olan lojistik regresyonda g\u00f6r\u00fcn\u00fcr. D\u00fczeltilmi\u015f olas\u0131l\u0131k oran\u0131, bir modeldeki […]<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[11],"tags":[],"class_list":["post-1528","post","type-post","status-publish","format-standard","hentry","category-rehber"],"yoast_head":"\n \u00d6rnek: D\u00fczeltilmi\u015f oran oranlar\u0131n\u0131n hesaplanmas\u0131<\/strong><\/span><\/h3>\n
![\"\"](\"https:\/\/statorials.org\/wp-content\/uploads\/2023\/08\/ajuste_odds_ratio1.png\")
<\/p>\n![\"D\u00fczeltilmi\u015f](\"https:\/\/statorials.org\/wp-content\/uploads\/2023\/08\/ajuste_odds_ratio2.png\")
<\/p>\n \u00d6zet: olas\u0131l\u0131k oran\u0131 ve d\u00fczeltilmi\u015f olas\u0131l\u0131k oran\u0131<\/strong><\/span><\/h3>\n
Ek kaynaklar<\/strong><\/span><\/h3>\n
R’de lojistik regresyon nas\u0131l ger\u00e7ekle\u015ftirilir
Python’da Lojistik Regresyon Nas\u0131l Ger\u00e7ekle\u015ftirilir<\/a><\/p>\n","protected":false},"excerpt":{"rendered":"