<\/p>\n
Genellikle makine \u00f6\u011frenimi algoritmalar\u0131n\u0131 veri k\u00fcmelerine uyarlad\u0131\u011f\u0131m\u0131zda, \u00f6ncelikle veri k\u00fcmesini bir e\u011fitim seti ve bir test seti olarak ikiye ay\u0131r\u0131r\u0131z.<\/span><\/p>\n R’de verileri e\u011fitim ve test k\u00fcmelerine b\u00f6lmenin \u00fc\u00e7 yayg\u0131n yolu vard\u0131r:<\/span><\/p>\n Y\u00f6ntem 1: Base R’yi kullan\u0131n<\/strong><\/span><\/p>\n Y\u00f6ntem 2: caTools paketini kullanma<\/strong><\/span><\/p>\n Y\u00f6ntem 3: dplyr paketini kullanma<\/strong><\/span><\/p>\n A\u015fa\u011f\u0131daki \u00f6rnekler, R’deki yerle\u015fik iris veri k\u00fcmesiyle<\/a> her y\u00f6ntemin pratikte nas\u0131l kullan\u0131laca\u011f\u0131n\u0131 g\u00f6sterir.<\/span><\/p>\n A\u015fa\u011f\u0131daki kod, sat\u0131rlar\u0131n %70’ini e\u011fitim seti ve geri kalan %30’unu test seti olarak kullanarak iris veri setini bir e\u011fitim ve test setine b\u00f6lmek i\u00e7in R taban\u0131n\u0131n nas\u0131l kullan\u0131laca\u011f\u0131n\u0131 g\u00f6sterir:<\/span><\/span><\/p>\n Sonu\u00e7tan \u015funu g\u00f6rebiliriz:<\/span><\/p>\n Orijinal veritaban\u0131 toplamda 150 sat\u0131ra sahip oldu\u011fundan e\u011fitim seti orijinal sat\u0131rlar\u0131n yakla\u015f\u0131k 106\/150 = %70,6’s\u0131n\u0131 i\u00e7ermektedir.<\/span><\/p>\n \u0130stersek e\u011fitim setinin ilk birka\u00e7 sat\u0131r\u0131n\u0131 da g\u00f6r\u00fcnt\u00fcleyebiliriz:<\/span><\/p>\n A\u015fa\u011f\u0131daki kod, sat\u0131rlar\u0131n %70’ini e\u011fitim seti ve geri kalan %30’unu test seti olarak kullanarak iris veri setini bir e\u011fitim ve test setine b\u00f6lmek i\u00e7in R’deki caTools<\/strong> paketinin nas\u0131l kullan\u0131laca\u011f\u0131n\u0131 g\u00f6sterir:<\/span><\/p>\n Sonu\u00e7tan \u015funu g\u00f6rebiliriz:<\/span><\/p>\n A\u015fa\u011f\u0131daki kod, sat\u0131rlar\u0131n %70’ini e\u011fitim seti ve geri kalan %30’unu test seti olarak kullanarak iris veri setini bir e\u011fitim ve test setine b\u00f6lmek i\u00e7in R’deki caTools<\/strong> paketinin nas\u0131l kullan\u0131laca\u011f\u0131n\u0131 g\u00f6sterir:<\/span><\/p>\n Sonu\u00e7tan \u015funu g\u00f6rebiliriz:<\/span><\/p>\n Bu e\u011fitim ve test setlerinin, olu\u015fturdu\u011fumuz ek bir “id” s\u00fctunu i\u00e7erdi\u011fini unutmay\u0131n.<\/span><\/p>\n Makine \u00f6\u011frenimi algoritman\u0131z\u0131 ayarlarken bu s\u00fctunu kullanmad\u0131\u011f\u0131n\u0131zdan (veya veri \u00e7er\u00e7evelerinden tamamen kald\u0131rmad\u0131\u011f\u0131n\u0131zdan) emin olun.<\/span><\/p>\n A\u015fa\u011f\u0131daki e\u011fitimlerde R’de di\u011fer yayg\u0131n i\u015flemlerin nas\u0131l ger\u00e7ekle\u015ftirilece\u011fi a\u00e7\u0131klanmaktad\u0131r:<\/span><\/p>\n R’de MSE nas\u0131l hesaplan\u0131r<\/a> Genellikle makine \u00f6\u011frenimi algoritmalar\u0131n\u0131 veri k\u00fcmelerine uyarlad\u0131\u011f\u0131m\u0131zda, \u00f6ncelikle veri k\u00fcmesini bir e\u011fitim seti ve bir test seti olarak ikiye ay\u0131r\u0131r\u0131z. R’de verileri e\u011fitim ve test k\u00fcmelerine b\u00f6lmenin \u00fc\u00e7 yayg\u0131n yolu vard\u0131r: Y\u00f6ntem 1: Base R’yi kullan\u0131n #make this example reproducible set. seeds (1) #use 70% of dataset as training set and 30% as test set […]<\/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-3044","post","type-post","status-publish","format-standard","hentry","category-rehber"],"yoast_head":"\n #make this example reproducible\n<\/span>set. seeds<\/span> (1)\n\n#use 70% of dataset as training set and 30% as test set\n<\/span>sample <- sample(c( TRUE<\/span> , FALSE<\/span> ), nrow(df), replace= TRUE<\/span> , prob=c( 0.7<\/span> , 0.3<\/span> ))\ntrain <- df[sample, ]\ntest <- df[!sample, ]<\/strong><\/pre>\n library<\/span> (caTools)<\/span>\n\n#make this example reproducible\n<\/span>set. seeds<\/span> (1)\n\n#use 70% of dataset as training set and 30% as test set\n<\/span>sample <- sample. split<\/span> (df$any_column_name, SplitRatio = 0.7<\/span> )\ntrain <- subset(df, sample == TRUE<\/span> )\ntest <- subset(df, sample == FALSE<\/span> )<\/strong><\/pre>\n library<\/span> (dplyr)<\/span>\n\n#make this example reproducible\n<\/span>set. seeds<\/span> (1)\n\n#create ID column\n<\/span>df$id <- 1:nrow(df)\n\n#use 70% of dataset as training set and 30% as test set<\/span>\ntrain <- df %>% dplyr::sample_frac( 0.70<\/span> )\ntest <- dplyr::anti_join(df, train, by = ' id<\/span> ')<\/strong><\/pre>\n \u00d6rnek 1: Base R’yi kullanarak verileri e\u011fitim ve test setlerine ay\u0131r\u0131n<\/strong><\/span><\/h3>\n
#load iris dataset\n<\/span>data(iris)\n\n#make this example reproducible\n<\/span>set. seeds<\/span> (1)\n\n#Use 70% of dataset as training set and remaining 30% as testing set\n<\/span>sample <- sample(c( TRUE<\/span> , FALSE<\/span> ), nrow(iris), replace= TRUE<\/span> , prob=c( 0.7<\/span> , 0.3<\/span> ))\ntrain <- iris[sample, ]\ntest <- iris[!sample, ]\n\n#view dimensions of training set\n<\/span>sun(train)\n\n[1] 106 5\n\n#view dimensions of test set\n<\/span>dim(test)\n\n[1] 44 5<\/strong><\/pre>\n\n
#view first few rows of training set\n<\/span>head(train)\n\n Sepal.Length Sepal.Width Petal.Length Petal.Width Species\n1 5.1 3.5 1.4 0.2 setosa\n2 4.9 3.0 1.4 0.2 setosa\n3 4.7 3.2 1.3 0.2 setosa\n5 5.0 3.6 1.4 0.2 setosa\n8 5.0 3.4 1.5 0.2 setosa\n9 4.4 2.9 1.4 0.2 setosa\n<\/strong><\/pre>\n \u00d6rnek 2: caTools’u kullanarak verileri e\u011fitim ve test setlerine ay\u0131r\u0131n<\/strong><\/span><\/h3>\n
library<\/span> (caTools)<\/span>\n\n#load iris dataset\n<\/span>data(iris)\n\n#make this example reproducible\n<\/span>set. seeds<\/span> (1)\n\n#Use 70% of dataset as training set and remaining 30% as testing set\n<\/span>sample <- sample. split<\/span> (iris$Species, SplitRatio = 0.7<\/span> )\ntrain <- subset(iris, sample == TRUE<\/span> )\ntest <- subset(iris, sample == FALSE<\/span> )\n\n#view dimensions of training set\n<\/span>sun(train)\n\n[1] 105 5\n\n#view dimensions of test set\n<\/span>dim(test)\n\n[1] 45 5<\/strong><\/pre>\n\n
\u00d6rnek 3: dplyr kullanarak verileri e\u011fitim ve test k\u00fcmelerine b\u00f6lme<\/strong><\/span><\/h3>\n
library<\/span> (dplyr)<\/span>\n\n#load iris dataset\n<\/span>data(iris)\n\n#make this example reproducible\n<\/span>set. seeds<\/span> (1)\n\n#create variable ID\n<\/span>iris$id <- 1:nrow(iris)\n\n#Use 70% of dataset as training set and remaining 30% as testing set<\/span> \ntrain <- iris %>% dplyr::sample_frac( 0.7<\/span> )\ntest <- dplyr::anti_join(iris, train, by = ' id<\/span> ')\n\n#view dimensions of training set\n<\/span>sun(train)\n\n[1] 105 6\n\n#view dimensions of test set\n<\/span>dim(test)\n\n[1] 45 6\n<\/strong><\/pre>\n\n
Ek kaynaklar<\/strong><\/span><\/h3>\n
R’de RMSE nas\u0131l hesaplan\u0131r<\/a>
R’de d\u00fczeltilmi\u015f R-kare nas\u0131l hesaplan\u0131r<\/a><\/p>\n","protected":false},"excerpt":{"rendered":"