<\/p>\n
Rand endeksi,<\/strong> iki farkl\u0131 k\u00fcmeleme y\u00f6ntemi aras\u0131ndaki sonu\u00e7lar\u0131n benzerli\u011fini kar\u015f\u0131la\u015ft\u0131rman\u0131n bir yoludur.<\/span><\/p>\n \u00c7o\u011funlukla R<\/em> olarak g\u00f6sterilen Rand endeksi \u015fu \u015fekilde hesaplan\u0131r:<\/span><\/p>\n R<\/em> = (a+b) \/ ( nC2<\/sub> )<\/sub><\/strong><\/span><\/p>\n Alt\u0131n:<\/span><\/p>\n Rand endeksi her zaman 0 ile 1 aras\u0131nda bir de\u011fer al\u0131r; burada:<\/span><\/p>\n A\u015fa\u011f\u0131daki \u00f6rnek, basit bir veri k\u00fcmesi i\u00e7in iki k\u00fcmeleme y\u00f6ntemi aras\u0131nda Rand dizininin nas\u0131l hesaplanaca\u011f\u0131n\u0131 g\u00f6sterir.<\/span><\/p>\n Be\u015f \u00f6\u011feden olu\u015fan a\u015fa\u011f\u0131daki veri setine sahip oldu\u011fumuzu varsayal\u0131m:<\/span><\/p>\n Ve her bir \u00f6\u011feyi a\u015fa\u011f\u0131daki k\u00fcmelere yerle\u015ftiren iki k\u00fcmeleme y\u00f6ntemi kulland\u0131\u011f\u0131m\u0131z\u0131 varsayal\u0131m:<\/span><\/p>\n Bu k\u00fcmeleme y\u00f6ntemleri aras\u0131nda Rand endeksini hesaplamak i\u00e7in \u00f6ncelikle be\u015f \u00f6\u011feli veri k\u00fcmesindeki t\u00fcm olas\u0131 s\u0131ras\u0131z \u00e7iftleri yazmam\u0131z gerekir:<\/span><\/p>\n 10 tane<\/strong> s\u0131ras\u0131z \u00e7ift vard\u0131r.<\/span><\/p>\n Daha sonra, her iki k\u00fcmeleme y\u00f6nteminde de ayn\u0131 k\u00fcmeye ait s\u0131ras\u0131z \u00e7iftlerin say\u0131s\u0131n\u0131 temsil eden a’y\u0131<\/strong> hesaplamam\u0131z gerekir:<\/span><\/p>\n Bu durumda a = 1<\/strong> olur.<\/span><\/p>\n Daha sonra, iki k\u00fcmeleme y\u00f6nteminde farkl\u0131 k\u00fcmelere ait s\u0131ras\u0131z \u00e7iftlerin say\u0131s\u0131n\u0131 temsil eden b’yi<\/strong> hesaplamam\u0131z gerekir:<\/span><\/p>\n Bu durumda b = 5<\/strong> olur.<\/span><\/p>\n Son olarak Rand indeksini \u015fu \u015fekilde hesaplayabiliriz:<\/span><\/p>\n Rand endeksi 0,6’d\u0131r<\/strong> .<\/span><\/p>\n R’deki iki k\u00fcmeleme y\u00f6ntemi aras\u0131ndaki Rand indeksini hesaplamak i\u00e7in fosil<\/strong> paketindeki Rand.index()<\/strong> fonksiyonunu kullanabiliriz:<\/span><\/p>\n Rand endeksi 0,6’d\u0131r<\/strong> . Bu bizim manuel olarak hesaplad\u0131\u011f\u0131m\u0131z de\u011fere kar\u015f\u0131l\u0131k geliyor.<\/span><\/p>\n \u0130ki k\u00fcme aras\u0131ndaki Rand indeksini hesaplamak i\u00e7in Python’da a\u015fa\u011f\u0131daki fonksiyonu tan\u0131mlayabiliriz:<\/span><\/p>\n Rand endeksi 0,6<\/strong> olarak \u00e7\u0131k\u0131yor. Bu, \u00f6nceki \u00f6rneklerde hesaplanan de\u011fere kar\u015f\u0131l\u0131k gelir.<\/span><\/p>\n K-Means K\u00fcmelenmesine Giri\u015f<\/a> Rand endeksi, iki farkl\u0131 k\u00fcmeleme y\u00f6ntemi aras\u0131ndaki sonu\u00e7lar\u0131n benzerli\u011fini kar\u015f\u0131la\u015ft\u0131rman\u0131n bir yoludur. \u00c7o\u011funlukla R olarak g\u00f6sterilen Rand endeksi \u015fu \u015fekilde hesaplan\u0131r: R = (a+b) \/ ( nC2 ) Alt\u0131n: a: iki k\u00fcmeleme y\u00f6ntemine g\u00f6re bir \u00f6\u011fe \u00e7iftinin ayn\u0131 k\u00fcmeye ait olma say\u0131s\u0131. b: iki k\u00fcmeleme y\u00f6ntemine g\u00f6re bir \u00f6\u011fe \u00e7iftinin fark k\u00fcmelerine ait olma say\u0131s\u0131. […]<\/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-1643","post","type-post","status-publish","format-standard","hentry","category-rehber"],"yoast_head":"\n\n
\n
\u00d6rnek: Rand endeksi nas\u0131l hesaplan\u0131r?<\/strong><\/span><\/h3>\n
\n
\n
\n
\n
\n
\n
R’de Rand endeksi nas\u0131l hesaplan\u0131r<\/strong><\/span><\/h3>\n
library<\/span> (fossil)\n\n#define clusters\n<\/span>method1 <- c(1, 1, 1, 2, 2)\nmethod2 <- c(1, 1, 2, 2, 3)\n\n#calculate Rand index between clustering methods\n<\/span>rand. index<\/span> (method1, method2)\n\n[1] 0.6\n<\/strong><\/span><\/pre>\n Python’da Rand Dizini Nas\u0131l Hesaplan\u0131r?<\/strong><\/span><\/h3>\n
import<\/span> numpy as<\/span> np\nfrom<\/span> scipy. special<\/span> import<\/span> comb\n\n#define Rand index function\n<\/span>def<\/span> rand_index<\/span> (actual, pred):\n\n tp_plus_fp = comb(np. bincount<\/span> (actual), 2). sum<\/span> ()\n tp_plus_fn = comb(np. bincount<\/span> (pred), 2). sum<\/span> ()\n A = np. c_<\/span> [(actual, pred)]\n tp = sum(comb( np.bincount<\/span> (A[A[:, 0] == i, 1]), 2) .sum<\/span> ()\n for<\/span> i in<\/span> set(actual))\n fp = tp_plus_fp - tp\n fn = tp_plus_fn - tp\n tn = comb(len(A), 2) - tp - fp - fn\n return<\/span> (tp + tn) \/ (tp + fp + fn + tn)\n\n#calculate Rand index\n<\/span>rand_index([1, 1, 1, 2, 2], [1, 1, 2, 2, 3])\n\n0.6<\/strong><\/pre>\n Ek kaynaklar<\/strong><\/span><\/h3>\n
K-Medoids k\u00fcmelemeye giri\u015f<\/a>
Hiyerar\u015fik k\u00fcmelemeye giri\u015f<\/p>\n","protected":false},"excerpt":{"rendered":"