<\/p>\n
Bir\u00e7ok istatistiksel test, g\u00f6zlemlerin<\/a> ba\u011f\u0131ms\u0131z oldu\u011funu varsayar. Bu, bir veri setindeki hi\u00e7bir g\u00f6zlemin birbiriyle ili\u015fkili olmad\u0131\u011f\u0131 veya herhangi bir \u015fekilde birbirini etkilemedi\u011fi anlam\u0131na gelir.<\/span><\/p>\n \u00d6rne\u011fin iki kedi t\u00fcr\u00fc aras\u0131nda ortalama a\u011f\u0131rl\u0131k fark\u0131 olup olmad\u0131\u011f\u0131n\u0131 test etmek istedi\u011fimizi varsayal\u0131m. A t\u00fcr\u00fcnden 10 kedinin ve B t\u00fcr\u00fcnden 10 kedinin a\u011f\u0131rl\u0131\u011f\u0131n\u0131 \u00f6l\u00e7seydik, kedi gruplar\u0131n\u0131n her birinin ayn\u0131 \u00e7\u00f6pten gelmesi durumunda ba\u011f\u0131ms\u0131zl\u0131k varsay\u0131m\u0131n\u0131 ihlal etmi\u015f olurduk.<\/span><\/p>\n A T\u00fcr\u00fcn\u00fcn ana kedisinin tamamen d\u00fc\u015f\u00fck kilolu yavrulara sahip olmas\u0131, B T\u00fcr\u00fcn\u00fcn ana kedisinin ise a\u011f\u0131r yavrulara sahip olmas\u0131 m\u00fcmk\u00fcnd\u00fcr. Bu bak\u0131mdan her bir numunenin g\u00f6zlemleri birbirinden ba\u011f\u0131ms\u0131z de\u011fildir.<\/span><\/p>\n Bu ba\u011f\u0131ms\u0131zl\u0131k varsay\u0131m\u0131n\u0131 yapan \u00fc\u00e7 yayg\u0131n istatistiksel test t\u00fcr\u00fc vard\u0131r:<\/span><\/p>\n 1.<\/strong> \u0130ki \u00f6rnekli t testi<\/a><\/span><\/p>\n 2.<\/strong> ANOVA (Varyans Analizi)<\/a><\/span><\/p>\n 3.<\/strong> Do\u011frusal regresyon<\/a><\/span><\/p>\n A\u015fa\u011f\u0131daki b\u00f6l\u00fcmlerde bu varsay\u0131m\u0131n neden<\/em> her bir test t\u00fcr\u00fc i\u00e7in yap\u0131ld\u0131\u011f\u0131n\u0131 ve bu varsay\u0131m\u0131n kar\u015f\u0131lan\u0131p kar\u015f\u0131lanmad\u0131\u011f\u0131n\u0131n nas\u0131l belirlenece\u011fini a\u00e7\u0131klayaca\u011f\u0131z.<\/span><\/p>\n \u0130ki \u00f6rneklem t testi,<\/strong> iki pop\u00fclasyonun ortalamalar\u0131n\u0131n e\u015fit olup olmad\u0131\u011f\u0131n\u0131 test etmek i\u00e7in kullan\u0131l\u0131r.<\/span><\/p>\n Varsay\u0131m:<\/strong> Bu test t\u00fcr\u00fc, her bir numune i\u00e7indeki<\/em> g\u00f6zlemlerin birbirinden ba\u011f\u0131ms\u0131z oldu\u011funu ve numuneler aras\u0131ndaki<\/em> g\u00f6zlemlerin de birbirinden ba\u011f\u0131ms\u0131z oldu\u011funu varsayar.<\/span><\/p>\n Bu hipotezi test edin: Bu<\/strong> hipotezi test etmenin en basit yolu, her g\u00f6zlemin her \u00f6rnekte yaln\u0131zca bir kez g\u00f6r\u00fcnd\u00fc\u011f\u00fcn\u00fc ve her \u00f6rnekteki g\u00f6zlemlerin rastgele \u00f6rneklemeyle topland\u0131\u011f\u0131n\u0131 do\u011frulamakt\u0131r.<\/span><\/p>\n ANOVA,<\/strong> \u00fc\u00e7 veya daha fazla ba\u011f\u0131ms\u0131z grubun ortalamalar\u0131 aras\u0131nda anlaml\u0131 bir fark olup olmad\u0131\u011f\u0131n\u0131 belirlemek i\u00e7in kullan\u0131l\u0131r.<\/span><\/p>\n Varsay\u0131m:<\/strong> ANOVA, her gruptaki g\u00f6zlemlerin birbirinden ba\u011f\u0131ms\u0131z oldu\u011funu ve gruplar i\u00e7indeki g\u00f6zlemlerin rastgele bir \u00f6rnekle elde edildi\u011fini varsayar.<\/span><\/p>\n Bu hipotezi test edin:<\/strong> T testine benzer \u015fekilde, bu hipotezi test etmenin en basit yolu, her g\u00f6zlemin her \u00f6rnekte yaln\u0131zca bir kez g\u00f6r\u00fcnd\u00fc\u011f\u00fcn\u00fc ve her \u00f6rnekteki g\u00f6zlemlerin rastgele \u00f6rneklemeyle topland\u0131\u011f\u0131n\u0131 do\u011frulamakt\u0131r.<\/span><\/p>\n Do\u011frusal regresyon,<\/strong> bir veya daha fazla yorday\u0131c\u0131 de\u011fi\u015fken ile bir yan\u0131t de\u011fi\u015fkeni<\/a> aras\u0131ndaki ili\u015fkiyi anlamak i\u00e7in kullan\u0131l\u0131r.<\/span><\/p>\n Varsay\u0131m:<\/strong> Do\u011frusal regresyon, tak\u0131lan modelin art\u0131klar\u0131n\u0131n<\/a> ba\u011f\u0131ms\u0131z oldu\u011funu varsayar.<\/span><\/p>\n Bu hipotezi test edin: Bu<\/strong> hipotezi test etmenin en kolay yolu, art\u0131klar\u0131n zamana kar\u015f\u0131 grafi\u011fi olan, art\u0131klar\u0131n zaman serisi grafi\u011fine bakmakt\u0131r. \u0130deal olarak, art\u0131k otokorelasyonlar\u0131n \u00e7o\u011fu, n’nin<\/em> \u00f6rnek boyutu oldu\u011fu n’nin<\/em> karek\u00f6k\u00fcnde yakla\u015f\u0131k olarak +\/- 2’de yer alan s\u0131f\u0131r \u00e7evresindeki %95 g\u00fcven bantlar\u0131 i\u00e7erisine d\u00fc\u015fmelidir. Ayr\u0131ca Durbin-Watson testini<\/a> kullanarak bu varsay\u0131m\u0131n kar\u015f\u0131lan\u0131p kar\u015f\u0131lanmad\u0131\u011f\u0131n\u0131 resmi olarak test edebilirsiniz.<\/span><\/p>\n Veri k\u00fcmelerinde ba\u011f\u0131ms\u0131z olmaman\u0131n \u00fc\u00e7 yayg\u0131n kayna\u011f\u0131 vard\u0131r:<\/span><\/p>\n 1. G\u00f6zlemler zamanla birlikte kapat\u0131l\u0131r.<\/strong><\/span><\/p>\n \u00d6rne\u011fin bir ara\u015ft\u0131rmac\u0131, arabalar\u0131n belirli bir yoldaki ortalama h\u0131z\u0131na ili\u015fkin veri toplayabilir. Ak\u015famlar\u0131 h\u0131zlar\u0131 takip etmeyi se\u00e7erse, ortalama h\u0131z\u0131n bekledi\u011finden \u00e7ok daha y\u00fcksek oldu\u011funu g\u00f6rebilir \u00e7\u00fcnk\u00fc her s\u00fcr\u00fcc\u00fc i\u015ften eve ko\u015fuyor.<\/span><\/p>\n Bu veriler her g\u00f6zlemin ba\u011f\u0131ms\u0131z oldu\u011fu varsay\u0131m\u0131n\u0131 ihlal etmektedir. Her g\u00f6zlem g\u00fcn\u00fcn ayn\u0131 saatinde g\u00f6zlemlendi\u011finden, her araban\u0131n h\u0131z\u0131n\u0131n benzer olmas\u0131 muhtemeldir.<\/span><\/p>\n 2. G\u00f6zlemler uzayda birbirine kapal\u0131d\u0131r.<\/strong><\/span><\/p>\n \u00d6rne\u011fin bir ara\u015ft\u0131rmac\u0131, uygun oldu\u011fu i\u00e7in ayn\u0131 y\u00fcksek gelirli mahallede ya\u015fayan insanlardan y\u0131ll\u0131k gelir verilerini toplayabilir.<\/span><\/p>\n Bu ba\u011flamda, veri \u00f6rne\u011fine dahil edilen ki\u015filerin hepsi birbirine yak\u0131n ya\u015fad\u0131klar\u0131 i\u00e7in benzer gelirlere sahip olmalar\u0131 muhtemeldir. Bu, her g\u00f6zlemin ba\u011f\u0131ms\u0131z oldu\u011fu varsay\u0131m\u0131n\u0131 ihlal etmektedir.<\/span><\/p>\n 3. G\u00f6zlemler ayn\u0131 veri k\u00fcmesinde birden \u00e7ok kez g\u00f6r\u00fcn\u00fcr.<\/strong><\/span><\/p>\n \u00d6rne\u011fin, bir ara\u015ft\u0131rmac\u0131n\u0131n 50 ki\u015fi hakk\u0131nda veri toplamas\u0131 gerekebilir, ancak bunu yapmak \u00e7ok daha kolay oldu\u011fu i\u00e7in bunun yerine 25 ki\u015fi hakk\u0131nda iki kez veri toplamaya karar verebilir.<\/span><\/p>\n Bu durum ba\u011f\u0131ms\u0131zl\u0131k varsay\u0131m\u0131n\u0131 ihlal etmektedir \u00e7\u00fcnk\u00fc veri setindeki her g\u00f6zlem kendisiyle ili\u015fkili olacakt\u0131r.<\/span><\/p>\n Ba\u011f\u0131ms\u0131zl\u0131k varsay\u0131m\u0131n\u0131 ihlal etmekten ka\u00e7\u0131nman\u0131n en basit yolu, bir pop\u00fclasyondan \u00f6rnek al\u0131rken basit rastgele \u00f6rnekleme<\/a> kullanmakt\u0131r.<\/span><\/p>\n Bu y\u00f6ntemi kullanarak, ilgilenilen pop\u00fclasyondaki<\/a> her bireyin \u00f6rne\u011fe dahil olma \u015fans\u0131 e\u015fit olur.<\/span><\/p>\n \u00d6rne\u011fin, ilgilendi\u011fimiz pop\u00fclasyon 10.000 birey i\u00e7eriyorsa, pop\u00fclasyondaki her bireye rastgele bir say\u0131 atayabilir ve ard\u0131ndan 40 rastgele say\u0131 se\u00e7mek i\u00e7in bir rastgele say\u0131 \u00fcreteci kullanabiliriz. Bu say\u0131lara uyan bireyler daha sonra numuneye dahil edilecektir.<\/span><\/p>\n Bu y\u00f6ntemi kullanarak birbirine \u00e7ok yak\u0131n olabilecek veya bir \u015fekilde akraba olabilecek iki ki\u015fiyi se\u00e7me \u015fans\u0131n\u0131 en aza indiriyoruz.<\/span><\/p>\n Bu, a\u015fa\u011f\u0131daki gibi di\u011fer \u00f6rnekleme y\u00f6ntemlerinin tam tersidir:<\/span><\/p>\n Rastgele \u00f6rnekleme y\u00f6ntemi kullanarak ba\u011f\u0131ms\u0131zl\u0131k varsay\u0131m\u0131n\u0131n ihlal edilme olas\u0131l\u0131\u011f\u0131n\u0131 en aza indirebiliriz.<\/span><\/p>\n T testinde form\u00fcle edilen d\u00f6rt hipotez<\/a> Bir\u00e7ok istatistiksel test, g\u00f6zlemlerin ba\u011f\u0131ms\u0131z oldu\u011funu varsayar. Bu, bir veri setindeki hi\u00e7bir g\u00f6zlemin birbiriyle ili\u015fkili olmad\u0131\u011f\u0131 veya herhangi bir \u015fekilde birbirini etkilemedi\u011fi anlam\u0131na gelir. \u00d6rne\u011fin iki kedi t\u00fcr\u00fc aras\u0131nda ortalama a\u011f\u0131rl\u0131k fark\u0131 olup olmad\u0131\u011f\u0131n\u0131 test etmek istedi\u011fimizi varsayal\u0131m. A t\u00fcr\u00fcnden 10 kedinin ve B t\u00fcr\u00fcnden 10 kedinin a\u011f\u0131rl\u0131\u011f\u0131n\u0131 \u00f6l\u00e7seydik, kedi gruplar\u0131n\u0131n her birinin ayn\u0131 \u00e7\u00f6pten […]<\/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-1529","post","type-post","status-publish","format-standard","hentry","category-rehber"],"yoast_head":"\n T-testlerinde ba\u011f\u0131ms\u0131zl\u0131k varsay\u0131m\u0131<\/strong><\/span><\/h3>\n
ANOVA’da ba\u011f\u0131ms\u0131zl\u0131k varsay\u0131m\u0131<\/strong><\/span><\/h3>\n
Varsay\u0131lan regresyonda ba\u011f\u0131ms\u0131zl\u0131k<\/strong><\/span><\/h3>\n
Ba\u011f\u0131ms\u0131z Olmaman\u0131n Ortak Kaynaklar\u0131<\/strong><\/span><\/h3>\n
Ba\u011f\u0131ms\u0131zl\u0131k varsay\u0131m\u0131n\u0131n ihlal edilmesinden nas\u0131l ka\u00e7\u0131n\u0131l\u0131r?<\/strong><\/span><\/h3>\n
\n
Ek kaynaklar<\/strong><\/span><\/h3>\n
Do\u011frusal regresyonun d\u00f6rt varsay\u0131m\u0131<\/a>
ANOVA’n\u0131n \u00fc\u00e7 hipotezi<\/a>
Temsili \u00f6rnek nedir ve neden \u00f6nemlidir?<\/a><\/p>\n","protected":false},"excerpt":{"rendered":"