1. E\u011fitim seti<\/strong> : modeli e\u011fitmek i\u00e7in kullan\u0131l\u0131r.<\/span><\/p>\n 2. Do\u011frulama seti<\/strong> : model parametrelerini optimize etmek i\u00e7in kullan\u0131l\u0131r.<\/span><\/p>\n 3. Test seti<\/strong> : Nihai model performans\u0131na ili\u015fkin tarafs\u0131z bir tahmin elde etmek i\u00e7in kullan\u0131l\u0131r.<\/span><\/p>\n A\u015fa\u011f\u0131daki \u015femada bu \u00fc\u00e7 farkl\u0131 veri seti t\u00fcr\u00fcn\u00fcn g\u00f6rsel bir a\u00e7\u0131klamas\u0131 sunulmaktad\u0131r:<\/span> <\/p>\n \u00d6\u011frencilerin kafas\u0131n\u0131 kar\u0131\u015ft\u0131ran noktalardan biri do\u011frulama seti ile test seti aras\u0131ndaki farkt\u0131r.<\/span><\/p>\n Basit\u00e7e s\u00f6ylemek gerekirse, do\u011frulama seti<\/strong> model parametrelerini optimize etmek i\u00e7in kullan\u0131l\u0131rken , test seti<\/strong> nihai modelin tarafs\u0131z bir tahminini sa\u011flamak i\u00e7in kullan\u0131l\u0131r.<\/span><\/p>\n Model g\u00f6r\u00fcnmeyen bir veri k\u00fcmesine uyguland\u0131\u011f\u0131nda, k-katl\u0131 \u00e7apraz do\u011frulamayla \u00f6l\u00e7\u00fclen hata oran\u0131n\u0131n ger\u00e7ek hata oran\u0131n\u0131 oldu\u011fundan d\u00fc\u015f\u00fck tahmin etme e\u011filiminde oldu\u011fu g\u00f6sterilebilir.<\/span><\/p>\n B\u00f6ylece, ger\u00e7ek d\u00fcnyada ger\u00e7ek hata oran\u0131n\u0131n ne olaca\u011f\u0131na dair tarafs\u0131z bir tahmin elde etmek i\u00e7in son modeli test<\/strong> setine uyarl\u0131yoruz.<\/span><\/p>\n A\u015fa\u011f\u0131daki \u00f6rnek, pratikte do\u011frulama seti ile test seti aras\u0131ndaki fark\u0131 g\u00f6stermektedir.<\/span><\/p>\n Diyelim ki bir emlak yat\u0131r\u0131mc\u0131s\u0131, belirli bir evin sat\u0131\u015f fiyat\u0131n\u0131 tahmin etmek i\u00e7in (1) yatak odas\u0131 say\u0131s\u0131n\u0131, (2) toplam metrekare say\u0131s\u0131n\u0131 ve (3) banyo say\u0131s\u0131n\u0131 kullanmak istiyor.<\/span><\/p>\n Diyelim ki elinde 10.000 ev hakk\u0131nda bu bilgilerin yer ald\u0131\u011f\u0131 bir veri seti var. \u0130lk olarak veri setini 8.000 evden olu\u015fan bir e\u011fitim seti ve 2.000 evden olu\u015fan bir test setine b\u00f6lecek:<\/span> <\/p>\n Daha sonra veri setine d\u00f6rt kez \u00e7oklu do\u011frusal regresyon modeli s\u0131\u011fd\u0131r\u0131lacakt\u0131r. Her seferinde e\u011fitim seti i\u00e7in 6.000 ev ve do\u011frulama seti i\u00e7in 2.000 ev kullan\u0131lacakt\u0131r.<\/span><\/p>\n Buna k-katl\u0131 \u00e7apraz do\u011frulama denir.<\/strong><\/span><\/p>\n E\u011fitim seti modeli e\u011fitmek i\u00e7in, do\u011frulama seti ise modelin performans\u0131n\u0131 de\u011ferlendirmek i\u00e7in kullan\u0131l\u0131r. Do\u011frulama seti i\u00e7in her seferinde 2.000 evden olu\u015fan farkl\u0131 bir grup kullan\u0131lacakt\u0131r.<\/span> <\/p>\n En d\u00fc\u015f\u00fck hataya sahip modeli belirlemek (yani veri k\u00fcmesine en iyi uyan modeli belirlemek) i\u00e7in bu k-katl\u0131 \u00e7apraz do\u011frulamay\u0131 birka\u00e7 farkl\u0131 regresyon modeli t\u00fcr\u00fc \u00fczerinde ger\u00e7ekle\u015ftirebilir.<\/span><\/p>\n Ancak en iyi modeli belirledikten sonra, modelin nihai performans\u0131na ili\u015fkin tarafs\u0131z bir tahmin elde etmek i\u00e7in ba\u015flang\u0131\u00e7ta sundu\u011fu 2.000 evlik test setini kullanacak.<\/span> <\/p>\n \u00d6rne\u011fin, ortalama mutlak hatas\u0131 8,345<\/strong> olan belirli bir regresyon modeli t\u00fcr\u00fcn\u00fc tan\u0131mlayabilir. Yani \u00f6ng\u00f6r\u00fclen konut fiyat\u0131 ile ger\u00e7ek konut fiyat\u0131 aras\u0131ndaki ortalama mutlak fark 8.345 dolard\u0131r.<\/span><\/p>\n Daha sonra bu kesin regresyon modelini hen\u00fcz kullan\u0131lmam\u0131\u015f 2.000 evden olu\u015fan test setine uydurabilir ve modelin ortalama mutlak hatas\u0131n\u0131n 8,847<\/strong> oldu\u011funu bulabilir.<\/span><\/p>\n Dolay\u0131s\u0131yla modelin ger\u00e7ek ortalama mutlak hatas\u0131n\u0131n tarafs\u0131z tahmini 8.847 dolard\u0131r.<\/span><\/p>\n K Katlamal\u0131 \u00c7apraz Do\u011frulama \u0130\u00e7in Basit Bir K\u0131lavuz<\/a> Bir makine \u00f6\u011frenimi algoritmas\u0131n\u0131 bir veri k\u00fcmesine uyarlad\u0131\u011f\u0131m\u0131zda, genellikle veri k\u00fcmesini \u00fc\u00e7 par\u00e7aya b\u00f6leriz: 1. E\u011fitim seti : modeli e\u011fitmek i\u00e7in kullan\u0131l\u0131r. 2. Do\u011frulama seti : model parametrelerini optimize etmek i\u00e7in kullan\u0131l\u0131r. 3. Test seti : Nihai model performans\u0131na ili\u015fkin tarafs\u0131z bir tahmin elde etmek i\u00e7in kullan\u0131l\u0131r. A\u015fa\u011f\u0131daki \u015femada bu \u00fc\u00e7 farkl\u0131 veri seti t\u00fcr\u00fcn\u00fcn […]<\/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-2214","post","type-post","status-publish","format-standard","hentry","category-rehber"],"yoast_head":"\n![\"\"](\"https:\/\/statorials.org\/wp-content\/uploads\/2023\/08\/test_val1.png\")
<\/p>\n \u00d6rnek: Do\u011frulama seti ile test seti aras\u0131ndaki fark\u0131 anlama<\/strong><\/span><\/h3>\n
![\"\"](\"https:\/\/statorials.org\/wp-content\/uploads\/2023\/08\/test_val2.png\")
<\/p>\n![\"\"](\"https:\/\/statorials.org\/wp-content\/uploads\/2023\/08\/test_val3.png\")
<\/p>\n![\"\"](\"https:\/\/statorials.org\/wp-content\/uploads\/2023\/08\/test_val4.png\")
<\/p>\n Ek kaynaklar<\/strong><\/span><\/h3>\n
Python’da K-Fold \u00e7apraz do\u011frulama nas\u0131l ger\u00e7ekle\u015ftirilir?<\/a>
R’de K-Fold \u00e7apraz do\u011frulama nas\u0131l ger\u00e7ekle\u015ftirilir?<\/a><\/p>\n","protected":false},"excerpt":{"rendered":"