<\/p>\n
Do\u011frusal regresyonun temel varsay\u0131mlar\u0131ndan<\/a> biri, art\u0131klar aras\u0131nda herhangi bir korelasyonun olmamas\u0131, yani art\u0131klar\u0131n ba\u011f\u0131ms\u0131z olmas\u0131d\u0131r.<\/span><\/p>\n Birinci dereceden otokorelasyonu test etmek i\u00e7in Durbin-Watson testi<\/a> yapabiliriz. Ancak otokorelasyonu daha y\u00fcksek derecelerde test etmek istiyorsak Breusch-Godfrey testi<\/strong> yapmam\u0131z gerekir.<\/span><\/p>\n Bu test a\u015fa\u011f\u0131daki varsay\u0131mlar\u0131<\/a> kullan\u0131r:<\/span><\/p>\n H 0<\/sub> (s\u0131f\u0131r hipotezi):<\/strong> p’den k\u00fc\u00e7\u00fck veya p’ye<\/em> e\u015fit bir otokorelasyon yoktur.<\/span><\/p>\n HA<\/sub> (alternatif hipotez):<\/strong> p’den k\u00fc\u00e7\u00fck veya p’ye<\/em> e\u015fit belirli bir d\u00fczeyde otokorelasyon vard\u0131r.<\/span><\/p>\n Test istatisti\u011fi, p<\/em> serbestlik derecesine sahip bir Ki-kare da\u011f\u0131l\u0131m\u0131n\u0131 takip eder.<\/span><\/p>\n Bu test istatisti\u011fine kar\u015f\u0131l\u0131k gelen p de\u011feri<\/a> belirli bir anlaml\u0131l\u0131k d\u00fczeyinin (\u00f6rne\u011fin 0,05) alt\u0131ndaysa, o zaman s\u0131f\u0131r hipotezini reddedebilir ve art\u0131klar aras\u0131nda belirli bir d\u00fc\u015f\u00fck d\u00fczeyde veya p’ye<\/em> e\u015fit bir otokorelasyonun var oldu\u011fu sonucuna varabiliriz.<\/span><\/p>\n Python’da Breusch-Godfrey testi ger\u00e7ekle\u015ftirmek i\u00e7in statsmodels<\/b> k\u00fct\u00fcphanesindeki acorr_breusch_godfrey()<\/strong> fonksiyonunu kullanabilirsiniz.<\/span><\/p>\n A\u015fa\u011f\u0131daki ad\u0131m ad\u0131m \u00f6rnek, Python’da Breusch-Godfrey testinin nas\u0131l ger\u00e7ekle\u015ftirilece\u011fini a\u00e7\u0131klamaktad\u0131r.<\/span><\/p>\n \u00d6ncelikle iki \u00f6ng\u00f6r\u00fcc\u00fc de\u011fi\u015fken (x1 ve x2) ve bir yan\u0131t de\u011fi\u015fkeni (y) i\u00e7eren bir veri k\u00fcmesi olu\u015ftural\u0131m.<\/span><\/p>\n Daha sonra, yorday\u0131c\u0131 de\u011fi\u015fkenler olarak x1 ve x2’yi ve yan\u0131t de\u011fi\u015fkeni<\/a> olarak y’yi kullanarak \u00e7oklu do\u011frusal regresyon modelini<\/a> yerle\u015ftirebiliriz.<\/span><\/p>\n Daha sonra p<\/em> d\u00fczeyindeki art\u0131klar aras\u0131ndaki otokorelasyonu test etmek i\u00e7in Breusch-Godfrey testini ger\u00e7ekle\u015ftirece\u011fiz. Bu \u00f6rnek i\u00e7in p<\/em> = 3’\u00fc se\u00e7ece\u011fiz.<\/span><\/p>\n \u00c7\u0131k\u0131\u015f\u0131n ilk de\u011feri test istatisti\u011fini, ikinci de\u011feri ise kar\u015f\u0131l\u0131k gelen p de\u011ferini temsil eder.<\/span><\/p>\n Sonu\u00e7tan \u015funlar\u0131 g\u00f6rebiliriz:<\/span><\/p>\n Bu p de\u011feri 0,05’ten k\u00fc\u00e7\u00fck oldu\u011fundan s\u0131f\u0131r hipotezini reddedebilir ve 3’ten k\u00fc\u00e7\u00fck veya ona e\u015fit olan s\u0131radaki art\u0131klar aras\u0131nda bir otokorelasyon oldu\u011fu sonucuna varabiliriz.<\/span><\/p>\n S\u0131f\u0131r hipotezini reddederseniz ve art\u0131klarda otokorelasyonun mevcut oldu\u011fu sonucuna var\u0131rsan\u0131z, yeterince ciddi oldu\u011funu d\u00fc\u015f\u00fcn\u00fcyorsan\u0131z bu sorunu d\u00fczeltmek i\u00e7in birka\u00e7 se\u00e7ene\u011finiz vard\u0131r:<\/span><\/p>\n Python’da Do\u011frusal Regresyon \u0130\u00e7in Tam Bir K\u0131lavuz<\/a> Do\u011frusal regresyonun temel varsay\u0131mlar\u0131ndan biri, art\u0131klar aras\u0131nda herhangi bir korelasyonun olmamas\u0131, yani art\u0131klar\u0131n ba\u011f\u0131ms\u0131z olmas\u0131d\u0131r. Birinci dereceden otokorelasyonu test etmek i\u00e7in Durbin-Watson testi yapabiliriz. Ancak otokorelasyonu daha y\u00fcksek derecelerde test etmek istiyorsak Breusch-Godfrey testi yapmam\u0131z gerekir. Bu test a\u015fa\u011f\u0131daki varsay\u0131mlar\u0131 kullan\u0131r: H 0 (s\u0131f\u0131r hipotezi): p’den k\u00fc\u00e7\u00fck veya p’ye e\u015fit bir otokorelasyon yoktur. HA (alternatif […]<\/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-1638","post","type-post","status-publish","format-standard","hentry","category-rehber"],"yoast_head":"\n 1. Ad\u0131m: Verileri olu\u015fturun<\/strong><\/span><\/h3>\n
import<\/span> pandas as<\/span> pd<\/span>\n\n#create dataset<\/span>\ndf = pd. DataFrame<\/span> ({' x1<\/span> ': [3, 4, 4, 5, 8, 9, 11, 13, 14, 16, 17, 20],\n ' x2<\/span> ': [7, 7, 8, 8, 12, 4, 5, 15, 9, 17, 19, 19],\n ' y<\/span> ': [24, 25, 25, 27, 29, 31, 34, 34, 39, 30, 40, 49]})\n\n#view first five rows of dataset\n<\/span>df. head<\/span> ()\n\n\tx1 x2 y\n0 3 7 24\n1 4 7 25\n2 4 8 25\n3 5 8 27\n4 8 12 29\n<\/strong><\/pre>\n Ad\u0131m 2: Bir regresyon modeli yerle\u015ftirin<\/strong><\/span><\/h3>\n
import<\/span> statsmodels. api<\/span> as<\/span> sm\n\n#define response variable\n<\/span>y = df[' y<\/span> ']\n\n#define predictor variables\n<\/span>x = df[[' x1<\/span> ', ' x2<\/span> ']]\n\n#add constant to predictor variables\n<\/span>x = sm. add_constant<\/span> (x)\n\n#fit linear regression model\n<\/span>model = sm. OLS<\/span> (y,x). fit<\/span> ()\n<\/strong><\/span><\/pre>\n Ad\u0131m 3: Breusch-Godfrey testini ger\u00e7ekle\u015ftirin<\/strong><\/span><\/h3>\n
import<\/span> statsmodels. stats<\/span> . diagnosis<\/span> as<\/span> dg\n<\/span>\n#perform Breusch-Godfrey test at order p<\/em> = 3\nprint<\/span> (dg. acorr_breusch_godfrey<\/span> (model, nlags= 3<\/span> ))\n\n(8.70314827, 0.0335094873, 5.27967224, 0.0403980576)\n<\/span><\/span><\/strong><\/span><\/pre>\n\n
Otokorelasyonla nas\u0131l ba\u015fa \u00e7\u0131k\u0131l\u0131r?<\/strong><\/span><\/h3>\n
\n
Ek kaynaklar<\/strong><\/span><\/h3>\n
Python’da Durbin-Watson testi nas\u0131l yap\u0131l\u0131r?<\/a>
Python’da Ljung-Box testi nas\u0131l yap\u0131l\u0131r<\/a><\/p>\n","protected":false},"excerpt":{"rendered":"