K-ortalama k\u00fcmelemesi, bir veri k\u00fcmesindeki her g\u00f6zlemi K<\/em> k\u00fcmesinden birine yerle\u015ftirdi\u011fimiz bir tekniktir.<\/span><\/p>\n Nihai ama\u00e7, her k\u00fcme i\u00e7indeki g\u00f6zlemlerin birbirine olduk\u00e7a benzer oldu\u011fu, farkl\u0131 k\u00fcmelerdeki g\u00f6zlemlerin ise birbirinden olduk\u00e7a farkl\u0131 oldu\u011fu K<\/em> k\u00fcmeye sahip olmakt\u0131r.<\/span><\/p>\n Uygulamada, K-arac\u0131 k\u00fcmelemesini ger\u00e7ekle\u015ftirmek i\u00e7in a\u015fa\u011f\u0131daki ad\u0131mlar\u0131 kullan\u0131r\u0131z:<\/span><\/p>\n 1. K<\/em> i\u00e7in bir de\u011fer se\u00e7in.<\/strong><\/span><\/p>\n 2. Her g\u00f6zlemi rastgele olarak 1’den K’ye<\/em> kadar bir ba\u015flang\u0131\u00e7 k\u00fcmesine atay\u0131n.<\/strong><\/span><\/p>\n 3. K\u00fcme atamalar\u0131n\u0131n de\u011fi\u015fmesi durana kadar a\u015fa\u011f\u0131daki prosed\u00fcr\u00fc uygulay\u0131n.<\/strong><\/span><\/p>\n A\u015fa\u011f\u0131daki ad\u0131m ad\u0131m \u00f6rnek, sklearn<\/strong> mod\u00fcl\u00fcndeki KMeans<\/strong> i\u015flevini kullanarak Python’da k-means k\u00fcmelemesinin nas\u0131l ger\u00e7ekle\u015ftirilece\u011fini g\u00f6sterir.<\/span><\/p>\n \u00d6ncelikle k-means k\u00fcmelemesini ger\u00e7ekle\u015ftirmek i\u00e7in ihtiya\u00e7 duyaca\u011f\u0131m\u0131z t\u00fcm mod\u00fclleri i\u00e7e aktaraca\u011f\u0131z:<\/span><\/p>\n Daha sonra 20 farkl\u0131 basketbol oyuncusu i\u00e7in a\u015fa\u011f\u0131daki \u00fc\u00e7 de\u011fi\u015fkeni i\u00e7eren bir DataFrame olu\u015fturaca\u011f\u0131z:<\/span><\/p>\n A\u015fa\u011f\u0131daki kod, bu pandalar\u0131n DataFrame’inin nas\u0131l olu\u015fturulaca\u011f\u0131n\u0131 g\u00f6sterir:<\/span><\/p>\n Benzer akt\u00f6rleri bu \u00fc\u00e7 \u00f6l\u00e7\u00fcme g\u00f6re grupland\u0131rmak i\u00e7in k-means k\u00fcmelemesini kullanaca\u011f\u0131z.<\/span><\/p>\n Daha sonra a\u015fa\u011f\u0131daki ad\u0131mlar\u0131 ger\u00e7ekle\u015ftirece\u011fiz:<\/span><\/span><\/p>\n A\u015fa\u011f\u0131daki kod bunun nas\u0131l yap\u0131laca\u011f\u0131n\u0131 g\u00f6sterir:<\/span><\/p>\n Not<\/strong> : k-means algoritmas\u0131n\u0131 ayarlarken her de\u011fi\u015fkenin ayn\u0131 \u00f6neme sahip olmas\u0131 i\u00e7in \u00f6l\u00e7eklendirme kullan\u0131yoruz. Aksi takdirde en geni\u015f aral\u0131klara sahip de\u011fi\u015fkenlerin etkisi \u00e7ok fazla olacakt\u0131r.<\/span><\/p>\n Python’da k-means k\u00fcmelemesini ger\u00e7ekle\u015ftirmek i\u00e7in sklearn<\/strong> mod\u00fcl\u00fcnden KMeans<\/strong> fonksiyonunu kullanabiliriz.<\/span><\/p>\n Bu i\u015flev a\u015fa\u011f\u0131daki temel s\u00f6zdizimini kullan\u0131r:<\/span><\/p>\n KMeans(init=’rastgele’, n_clusters=8, n_init=10, random_state=Yok)<\/strong><\/span><\/p>\n Alt\u0131n:<\/span><\/p>\n Bu i\u015flevin en \u00f6nemli arg\u00fcman\u0131, g\u00f6zlemlerin ka\u00e7 k\u00fcmeye yerle\u015ftirilece\u011fini belirten n_clusters’t\u0131r.<\/span><\/p>\n Ancak, ka\u00e7 k\u00fcmenin optimal oldu\u011funu \u00f6nceden bilmiyoruz, bu nedenle k\u00fcme say\u0131s\u0131n\u0131 ve modelin SSE’sini (hatalar\u0131n karelerinin toplam\u0131) g\u00f6steren bir grafik olu\u015fturmam\u0131z gerekiyor.<\/span><\/p>\n Tipik olarak, bu t\u00fcr bir \u00e7izim olu\u015fturdu\u011fumuzda, karelerin toplam\u0131n\u0131n “b\u00fck\u00fclmeye” veya d\u00fczle\u015fmeye ba\u015flad\u0131\u011f\u0131 bir “diz” arar\u0131z. Bu genellikle optimum k\u00fcme say\u0131s\u0131d\u0131r.<\/span><\/p>\n A\u015fa\u011f\u0131daki kod, x ekseninde k\u00fcme say\u0131s\u0131n\u0131 ve y ekseninde SSE’yi g\u00f6r\u00fcnt\u00fcleyen bu t\u00fcr grafi\u011fin nas\u0131l olu\u015fturulaca\u011f\u0131n\u0131 g\u00f6sterir:<\/span> <\/p>\n Bu grafikte k = 3 k\u00fcmede<\/strong> b\u00fck\u00fclme veya “diz” oldu\u011fu g\u00f6r\u00fclmektedir.<\/span><\/p>\n Dolay\u0131s\u0131yla bir sonraki ad\u0131mda k-ortalama k\u00fcmeleme modelimizi yerle\u015ftirirken 3 k\u00fcme kullanaca\u011f\u0131z.<\/span><\/p>\n Not<\/strong> : Ger\u00e7ek d\u00fcnyada, kullan\u0131lacak k\u00fcme say\u0131s\u0131n\u0131 se\u00e7mek i\u00e7in bu plan ve etki alan\u0131 uzmanl\u0131\u011f\u0131n\u0131n bir kombinasyonunun kullan\u0131lmas\u0131 \u00f6nerilir.<\/span><\/p>\n A\u015fa\u011f\u0131daki kod, k<\/em> \/3 i\u00e7in en uygun de\u011feri kullanarak veri k\u00fcmesinde k-ortalamal\u0131 k\u00fcmelemenin nas\u0131l ger\u00e7ekle\u015ftirilece\u011fini g\u00f6sterir:<\/span><\/p>\n Ortaya \u00e7\u0131kan tablo, DataFrame’deki her g\u00f6zlem i\u00e7in k\u00fcme atamalar\u0131n\u0131 g\u00f6sterir.<\/span><\/p>\n Bu sonu\u00e7lar\u0131n yorumlanmas\u0131n\u0131 kolayla\u015ft\u0131rmak i\u00e7in DataFrame’e her oyuncunun k\u00fcme atamas\u0131n\u0131 g\u00f6steren bir s\u00fctun ekleyebiliriz:<\/span><\/p>\n K\u00fcme<\/strong> s\u00fctunu, her oyuncunun atand\u0131\u011f\u0131 k\u00fcme numaras\u0131n\u0131 (0, 1 veya 2) i\u00e7erir.<\/span><\/p>\n Ayn\u0131 k\u00fcmeye ait oyuncular say\u0131<\/strong> , asist<\/strong> ve ribaund<\/strong> s\u00fctunlar\u0131nda yakla\u015f\u0131k olarak benzer de\u011ferlere sahiptir.<\/span><\/p>\n Not<\/strong> : Sklearn’in<\/strong> KMeans<\/strong> i\u015flevine ili\u015fkin t\u00fcm belgeleri burada<\/a> bulabilirsiniz.<\/span><\/p>\n A\u015fa\u011f\u0131daki e\u011fitimlerde Python’da di\u011fer genel g\u00f6revlerin nas\u0131l ger\u00e7ekle\u015ftirilece\u011fi a\u00e7\u0131klanmaktad\u0131r:<\/span><\/p>\n Python’da do\u011frusal regresyon nas\u0131l ger\u00e7ekle\u015ftirilir<\/a> Makine \u00f6\u011freniminde en yayg\u0131n k\u00fcmeleme algoritmalar\u0131ndan biri k-ortalama k\u00fcmeleme olarak bilinir. K-ortalama k\u00fcmelemesi, bir veri k\u00fcmesindeki her g\u00f6zlemi K k\u00fcmesinden birine yerle\u015ftirdi\u011fimiz bir tekniktir. Nihai ama\u00e7, her k\u00fcme i\u00e7indeki g\u00f6zlemlerin birbirine olduk\u00e7a benzer oldu\u011fu, farkl\u0131 k\u00fcmelerdeki g\u00f6zlemlerin ise birbirinden olduk\u00e7a farkl\u0131 oldu\u011fu K k\u00fcmeye sahip olmakt\u0131r. Uygulamada, K-arac\u0131 k\u00fcmelemesini ger\u00e7ekle\u015ftirmek i\u00e7in a\u015fa\u011f\u0131daki ad\u0131mlar\u0131 kullan\u0131r\u0131z: 1. […]<\/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-3580","post","type-post","status-publish","format-standard","hentry","category-rehber"],"yoast_head":"\n\n
\n
Ad\u0131m 1: Gerekli mod\u00fclleri i\u00e7e aktar\u0131n<\/strong><\/span><\/h2>\n
import<\/span> pandas as<\/span> pd\nimport<\/span> numpy as<\/span> np\nimport<\/span> matplotlib. pyplot<\/span> as<\/span> plt\nfrom<\/span> sklearn. cluster<\/span> import<\/span> KMeans\nfrom<\/span> sklearn. preprocessing<\/span> import<\/span> StandardScaler<\/strong><\/span><\/pre>\n Ad\u0131m 2: DataFrame’i olu\u015fturun<\/strong><\/span><\/h2>\n
\n
#createDataFrame\n<\/span>df = pd. DataFrame<\/span> ({' points<\/span> ': [18, np.nan, 19, 14, 14, 11, 20, 28, 30, 31,\n 35, 33, 29, 25, 25, 27, 29, 30, 19, 23],\n ' assists<\/span> ': [3, 3, 4, 5, 4, 7, 8, 7, 6, 9, 12, 14,\n np.nan, 9, 4, 3, 4, 12, 15, 11],\n ' rebounds<\/span> ': [15, 14, 14, 10, 8, 14, 13, 9, 5, 4,\n 11, 6, 5, 5, 3, 8, 12, 7, 6, 5]})\n\n#view first five rows of DataFrame\n<\/span>print<\/span> ( df.head<\/span> ())\n\n points assists rebounds\n0 18.0 3.0 15\n1 NaN 3.0 14\n2 19.0 4.0 14\n3 14.0 5.0 10\n4 14.0 4.0 8\n<\/strong><\/span><\/pre>\n 3. Ad\u0131m: DataFrame’i temizleyin ve haz\u0131rlay\u0131n<\/strong><\/span><\/h2>\n
\n
#drop rows with NA values in any columns\n<\/span>df = df. dropna<\/span> ()\n\n#create scaled DataFrame where each variable has mean of 0 and standard dev of 1\n<\/span>scaled_df = StandardScaler(). fit_transform<\/span> (df)\n\n#view first five rows of scaled DataFrame<\/span>\nprint<\/span> (scaled_df[:5])\n\n[[-0.86660275 -1.22683918 1.72722524]\n [-0.72081911 -0.96077767 1.45687694]\n [-1.44973731 -0.69471616 0.37548375]\n [-1.44973731 -0.96077767 -0.16521285]\n [-1.88708823 -0.16259314 1.45687694]]<\/strong><\/pre>\n Ad\u0131m 4: Optimum K\u00fcme Say\u0131s\u0131n\u0131 Bulun<\/strong><\/span><\/h2>\n
\n
#initialize kmeans parameters\n<\/span>kmeans_kwargs = {\n\" init<\/span> \": \" random<\/span> \",\n\" n_init<\/span> \": 10,\n\" random_state<\/span> \": 1,\n}\n\n#create list to hold SSE values for each k\n<\/span>sse = []\nfor<\/span> k in<\/span> range(1, 11):\n kmeans = KMeans(n_clusters=k, **<\/span> kmeans_kwargs)\n kmeans. fit<\/span> (scaled_df)\n sse. append<\/span> (kmeans.inertia_)\n\n#visualize results\n<\/span>plt. plot<\/span> (range(1, 11), sse)\nplt. xticks<\/span> (range(1, 11))\nplt. xlabel<\/span> (\" Number of Clusters<\/span> \")\nplt. ylabel<\/span> (\u201c SSE<\/span> \u201d)\nplt. show<\/span> ()<\/strong> <\/pre>\n![\"\"](\"https:\/\/statorials.org\/wp-content\/uploads\/2023\/08\/kmmoyenne1.jpg\")
<\/p>\n Ad\u0131m 5: Optimal K<\/em> ile K-Means K\u00fcmelemesini Ger\u00e7ekle\u015ftirin<\/strong><\/span><\/h2>\n
#instantiate the k-means class, using optimal number of clusters\n<\/span>kmeans = KMeans(init=\" random<\/span> \", n_clusters= 3<\/span> , n_init= 10<\/span> , random_state= 1<\/span> )\n\n#fit k-means algorithm to data\n<\/span>kmeans. fit<\/span> (scaled_df)\n\n#view cluster assignments for each observation\n<\/span>kmeans. labels_\n\n<\/span>array([1, 1, 1, 1, 1, 1, 2, 2, 0, 0, 0, 0, 2, 2, 2, 0, 0, 0]) \n<\/span><\/span><\/strong><\/pre>\n #append cluster assingments to original DataFrame\n<\/span>df[' cluster<\/span> '] = kmeans. labels_<\/span>\n\n#view updated DataFrame\n<\/span>print<\/span> (df)\n\n<\/span><\/span>points assists rebounds cluster\n0 18.0 3.0 15 1\n2 19.0 4.0 14 1\n3 14.0 5.0 10 1\n4 14.0 4.0 8 1\n5 11.0 7.0 14 1\n6 20.0 8.0 13 1\n7 28.0 7.0 9 2\n8 30.0 6.0 5 2\n9 31.0 9.0 4 0\n10 35.0 12.0 11 0\n11 33.0 14.0 6 0\n13 25.0 9.0 5 0\n14 25.0 4.0 3 2\n15 27.0 3.0 8 2\n16 29.0 4.0 12 2\n17 30.0 12.0 7 0\n18 19.0 15.0 6 0\n19 23.0 11.0 5 0\n<\/strong><\/pre>\n Ek kaynaklar<\/strong><\/span><\/h2>\n
Python’da Lojistik Regresyon Nas\u0131l Ger\u00e7ekle\u015ftirilir<\/a>
Python’da K-Fold \u00e7apraz do\u011frulama nas\u0131l ger\u00e7ekle\u015ftirilir<\/a><\/p>\n","protected":false},"excerpt":{"rendered":"