Benzer \u015fekilde, bir g\u00f6zlem, veri k\u00fcmesindeki di\u011fer g\u00f6zlemlerle kar\u015f\u0131la\u015ft\u0131r\u0131ld\u0131\u011f\u0131nda \u00e7ok daha ekstrem tahmin de\u011fi\u015fkenleri i\u00e7in bir veya daha fazla de\u011fere sahipse, y\u00fcksek kald\u0131ra\u00e7<\/strong> olarak kabul edilir.<\/span><\/p>\n Herhangi bir analiz t\u00fcr\u00fcnde ilk ad\u0131mlardan biri, belirli bir modelin sonu\u00e7lar\u0131 \u00fczerinde b\u00fcy\u00fck bir etkiye sahip olabilece\u011finden, y\u00fcksek etkiye sahip g\u00f6zlemlere daha yak\u0131ndan bakmakt\u0131r.<\/span><\/p>\n Bu e\u011fitimde, R’deki bir modeldeki her g\u00f6zlem i\u00e7in kald\u0131rac\u0131n nas\u0131l hesaplanaca\u011f\u0131 ve g\u00f6rselle\u015ftirilece\u011fine ili\u015fkin ad\u0131m ad\u0131m bir \u00f6rnek g\u00f6sterilmektedir.<\/span><\/p>\n \u0130lk olarak, R’de yerle\u015fik mtcars<\/strong> veri k\u00fcmesini kullanarak \u00e7oklu do\u011frusal regresyon modeli<\/a> olu\u015fturaca\u011f\u0131z:<\/span><\/p>\n Daha sonra, modeldeki her g\u00f6zlemin kald\u0131rac\u0131n\u0131 hesaplamak i\u00e7in hatvalues()<\/strong> fonksiyonunu kullanaca\u011f\u0131z:<\/span><\/p>\n Genellikle kald\u0131ra\u00e7 de\u011feri 2’den b\u00fcy\u00fck olan g\u00f6zlemlere daha yak\u0131ndan bakar\u0131z.<\/span><\/p>\n Bunu yapman\u0131n basit bir yolu, g\u00f6zlemleri kald\u0131ra\u00e7 de\u011ferlerine g\u00f6re azalan d\u00fczende s\u0131ralamakt\u0131r:<\/span><\/p>\n En y\u00fcksek kald\u0131ra\u00e7 de\u011ferinin 0,4966<\/strong> oldu\u011funu g\u00f6rebiliyoruz. Bu say\u0131 2’den b\u00fcy\u00fck olmad\u0131\u011f\u0131ndan veri setimizdeki g\u00f6zlemlerin hi\u00e7birinin y\u00fcksek kald\u0131raca sahip olmad\u0131\u011f\u0131n\u0131 biliyoruz.<\/span><\/p>\n Son olarak, her g\u00f6zlemin etkisini g\u00f6rselle\u015ftirmek i\u00e7in h\u0131zl\u0131 bir grafik olu\u015fturabiliriz:<\/span> <\/p>\n X ekseni veri k\u00fcmesindeki her g\u00f6zlemin indeksini g\u00f6r\u00fcnt\u00fcler ve y de\u011feri ise her g\u00f6zlem i\u00e7in kar\u015f\u0131l\u0131k gelen kald\u0131ra\u00e7 istatisti\u011fini g\u00f6r\u00fcnt\u00fcler.<\/span><\/p>\n R’de basit do\u011frusal regresyon nas\u0131l ger\u00e7ekle\u015ftirilir<\/a> \u0130statistikte, bir g\u00f6zlemin yan\u0131t de\u011fi\u015fkenine ili\u015fkin de\u011feri veri k\u00fcmesindeki di\u011fer g\u00f6zlemlerden \u00e7ok daha b\u00fcy\u00fckse, bu g\u00f6zlem ayk\u0131r\u0131 de\u011fer olarak kabul edilir. Benzer \u015fekilde, bir g\u00f6zlem, veri k\u00fcmesindeki di\u011fer g\u00f6zlemlerle kar\u015f\u0131la\u015ft\u0131r\u0131ld\u0131\u011f\u0131nda \u00e7ok daha ekstrem tahmin de\u011fi\u015fkenleri i\u00e7in bir veya daha fazla de\u011fere sahipse, y\u00fcksek kald\u0131ra\u00e7 olarak kabul edilir. Herhangi bir analiz t\u00fcr\u00fcnde ilk ad\u0131mlardan biri, belirli […]<\/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-1274","post","type-post","status-publish","format-standard","hentry","category-rehber"],"yoast_head":"\n 1. Ad\u0131m: Bir regresyon modeli olu\u015fturun<\/strong><\/span><\/h3>\n
#load the dataset<\/span>\ndata(mtcars)\n\n#fit a regression model<\/span>\nmodel <- lm(mpg~disp+hp, data=mtcars)\n\n#view model summary\n<\/span>summary(model)\n\nCoefficients:\n Estimate Std. Error t value Pr(>|t|) \n(Intercept) 30.735904 1.331566 23.083 < 2nd-16 ***\navailable -0.030346 0.007405 -4.098 0.000306 ***\nhp -0.024840 0.013385 -1.856 0.073679 . \n---\nSignificant. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1\n\nResidual standard error: 3.127 on 29 degrees of freedom\nMultiple R-squared: 0.7482, Adjusted R-squared: 0.7309 \nF-statistic: 43.09 on 2 and 29 DF, p-value: 2.062e-09\n<\/strong><\/pre>\n Ad\u0131m 2: Her g\u00f6zlem i\u00e7in kald\u0131rac\u0131 hesaplay\u0131n<\/strong><\/span><\/h3>\n
#calculate leverage for each observation in the model\n<\/span>hats <- as<\/span> . data<\/span> . frame<\/span> (hatvalues(model))\n\n#display leverage stats for each observation\n<\/span>hats\n\n hatvalues(model)\nMazda RX4 0.04235795\nMazda RX4 Wag 0.04235795\nDatsun 710 0.06287776\nHornet 4 Drive 0.07614472\nHornet Sportabout 0.08097817\nValiant 0.05945972\nDuster 360 0.09828955\nMerc 240D 0.08816960\nMerc 230 0.05102253\nMerc 280 0.03990060\nMerc 280C 0.03990060\nMerc 450SE 0.03890159\nMerc 450SL 0.03890159\nMerc 450SLC 0.03890159\nCadillac Fleetwood 0.19443875\nLincoln Continental 0.16042361\nChrysler Imperial 0.12447530\nFiat 128 0.08346304\nHonda Civic 0.09493784\nToyota Corolla 0.08732818\nToyota Corona 0.05697867\nDodge Challenger 0.06954069\nAMC Javelin 0.05767659\nCamaro Z28 0.10011654\nPontiac Firebird 0.12979822\nFiat X1-9 0.08334018\nPorsche 914-2 0.05785170\nLotus Europa 0.08193899\nFord Pantera L 0.13831817\nFerrari Dino 0.12608583\nMaserati Bora 0.49663919\nVolvo 142E 0.05848459\n<\/strong><\/pre>\n #sort observations by leverage, descending<\/span>\nhats[ order<\/span> (-hats[' hatvalues(model)<\/span> ']), ]\n\n [1] 0.49663919 0.19443875 0.16042361 0.13831817 0.12979822 0.12608583\n [7] 0.12447530 0.10011654 0.09828955 0.09493784 0.08816960 0.08732818\n[13] 0.08346304 0.08334018 0.08193899 0.08097817 0.07614472 0.06954069\n[19] 0.06287776 0.05945972 0.05848459 0.05785170 0.05767659 0.05697867\n[25] 0.05102253 0.04235795 0.04235795 0.03990060 0.03990060 0.03890159\n[31] 0.03890159 0.03890159\n<\/strong><\/pre>\n 3. Ad\u0131m: Her g\u00f6zlemin etkisini g\u00f6rselle\u015ftirin<\/strong><\/span><\/h3>\n
#plot leverage values for each observation<\/span>\nplot(hatvalues(model), type = ' h<\/span> ')\n<\/strong><\/pre>\n![\"R'de](\"https:\/\/statorials.org\/wp-content\/uploads\/2023\/08\/effet-de-levier1.png\")
<\/p>\n Ek kaynaklar<\/strong><\/span><\/h3>\n
R’de \u00e7oklu do\u011frusal regresyon nas\u0131l ger\u00e7ekle\u015ftirilir<\/a>
R’de art\u0131k arsa nas\u0131l olu\u015fturulur<\/a><\/p>\n","protected":false},"excerpt":{"rendered":"