{"id":2257,"date":"2023-07-23T01:30:54","date_gmt":"2023-07-23T01:30:54","guid":{"rendered":"https:\/\/statorials.org\/tr\/zayif-uyum-siralamasindan-tahmin-yaniltici-olabilir\/"},"modified":"2023-07-23T01:30:54","modified_gmt":"2023-07-23T01:30:54","slug":"zayif-uyum-siralamasindan-tahmin-yaniltici-olabilir","status":"publish","type":"post","link":"https:\/\/statorials.org\/tr\/zayif-uyum-siralamasindan-tahmin-yaniltici-olabilir\/","title":{"rendered":"Nas\u0131l d\u00fczeltilir? k\u00f6t\u00fc s\u0131ralama ayarlamas\u0131n\u0131 tahmin etmek yan\u0131lt\u0131c\u0131 olabilir"},"content":{"rendered":"

<\/p>\n

\n

R’de kar\u015f\u0131la\u015fabilece\u011finiz yayg\u0131n bir uyar\u0131:<\/span><\/p>\n

 Warning message:\nIn predict.lm(model, df):\n  prediction from a rank-deficient fit may be misleading\n<\/strong><\/span><\/pre>\n
 Bu uyar\u0131 iki nedenden dolay\u0131 ortaya \u00e7\u0131kabilir:<\/span><\/p>\n
 Sebep 1<\/strong> : \u0130ki yorday\u0131c\u0131 de\u011fi\u015fken m\u00fckemmel bir \u015fekilde ili\u015fkilidir.<\/span><\/p>\n
 Sebep 2<\/strong> : Veri k\u00fcmesindeki g\u00f6zlemlerden daha fazla model parametreniz var.<\/span><\/p>\n
 A\u015fa\u011f\u0131daki \u00f6rnekler her sorunun pratikte nas\u0131l ortaya \u00e7\u0131kabilece\u011fini g\u00f6stermektedir.<\/span><\/p>\n
 Sebep #1: \u0130ki yorday\u0131c\u0131 de\u011fi\u015fken m\u00fckemmel bir \u015fekilde ili\u015fkilidir<\/strong><\/span><\/h3>\n
 A\u015fa\u011f\u0131daki \u00e7oklu do\u011frusal regresyon modelini R’ye uydurdu\u011fumuzu ve onu tahminlerde bulunmak i\u00e7in kullanmaya \u00e7al\u0131\u015ft\u0131\u011f\u0131m\u0131z\u0131 varsayal\u0131m:<\/span><\/p>\n
 #create data frame\n<\/span>df <- data. frame<\/span> (x1=c(1, 2, 3, 4),\n                 x2=c(2, 4, 6, 8),\n                 y=c(6, 10, 19, 26))\n\n#fit multiple linear regression model\n<\/span>model <- lm(y~x1+x2, data=df)\n\n#use model to make predictions\n<\/span>predict(model, df)\n\n   1 2 3 4 \n 4.9 11.8 18.7 25.6 \nWarning message:\nIn predict.lm(model, df):\n  prediction from a rank-deficient fit may be misleading<\/span><\/span><\/strong><\/pre>\n
 Tahmin edici de\u011fi\u015fkenler x1 ve x2 m\u00fckemmel bir \u015fekilde ili\u015fkili<\/strong> oldu\u011fu i\u00e7in bir uyar\u0131 mesaj\u0131 al\u0131yoruz.<\/span><\/p>\n
 X2 de\u011ferlerinin basit\u00e7e x1 de\u011ferlerinin ikiyle \u00e7arp\u0131m\u0131na e\u015fit oldu\u011funu unutmay\u0131n. Bu m\u00fckemmel \u00e7oklu ba\u011flant\u0131n\u0131n<\/a> bir \u00f6rne\u011fidir.<\/span><\/p>\n
 Bu, x1 ve x2’nin regresyon modelinde benzersiz veya ba\u011f\u0131ms\u0131z bilgi sa\u011flamad\u0131\u011f\u0131 anlam\u0131na gelir; bu da modelin yerle\u015ftirilmesinde ve yorumlanmas\u0131nda sorunlara neden olur.<\/span><\/p>\n
 Bu sorunu \u00e7\u00f6zmenin en kolay yolu, tahmin de\u011fi\u015fkenlerinden birini modelden \u00e7\u0131karmakt\u0131r \u00e7\u00fcnk\u00fc modelde her iki tahmin de\u011fi\u015fkeninin de bulunmas\u0131 gereksizdir.<\/span><\/p>\n
 Sebep #2: G\u00f6zlemlerden daha fazla model parametresi var<\/strong><\/span><\/h3>\n
 A\u015fa\u011f\u0131daki \u00e7oklu do\u011frusal regresyon modelini R’ye uydurdu\u011fumuzu ve onu tahminlerde bulunmak i\u00e7in kullanmaya \u00e7al\u0131\u015ft\u0131\u011f\u0131m\u0131z\u0131 varsayal\u0131m:<\/span><\/p>\n
 #create data frame\n<\/span>df <- data. frame<\/span> (x1=c(1, 2, 3, 4),\n                 x2=c(3, 3, 8, 12),\n                 x3=c(4, 6, 3, 11),\n                 y=c(6, 10, 19, 26))\n\n#fit multiple linear regression model\n<\/span>model <- lm(y~x1*x2*x3, data=df)\n\n#use model to make predictions\n<\/span>predict(model, df)\n\n 1 2 3 4 \n 6 10 19 26 \nWarning message:\nIn predict.lm(model, df):\n  prediction from a rank-deficient fit may be misleading\n<\/span><\/span><\/strong><\/pre>\n
 Yedi toplam model katsay\u0131s\u0131na sahip bir regresyon modelini uydurmaya \u00e7al\u0131\u015ft\u0131\u011f\u0131m\u0131z i\u00e7in bir uyar\u0131 mesaj\u0131 al\u0131yoruz:<\/span><\/p>\n