<\/p>\n
Makine \u00f6\u011freniminde s\u0131n\u0131fland\u0131rma modellerini<\/a> kulland\u0131\u011f\u0131m\u0131zda modelin kalitesini de\u011ferlendirmek i\u00e7in \u00fc\u00e7 ortak \u00f6l\u00e7\u00fcm kullan\u0131r\u0131z:<\/span><\/p>\n 1. Do\u011fruluk<\/strong> : Toplam pozitif tahminlere k\u0131yasla do\u011fru pozitif tahminlerin y\u00fczdesi.<\/span><\/p>\n 2. Hat\u0131rlama<\/strong> : Toplam fiili pozitif tahminlere k\u0131yasla do\u011fru pozitif tahminlerin y\u00fczdesi.<\/span><\/p>\n 3. F1 Puan\u0131<\/strong> : Hassasiyet ve hat\u0131rlaman\u0131n a\u011f\u0131rl\u0131kl\u0131 harmonik ortalamas\u0131. Model 1’e ne kadar yak\u0131nsa model o kadar iyidir.<\/span><\/p>\n Bu \u00fc\u00e7 \u00f6l\u00e7\u00fcm\u00fc kullanarak, belirli bir s\u0131n\u0131fland\u0131rma modelinin belirli yan\u0131t de\u011fi\u015fkenlerine<\/a> ili\u015fkin sonu\u00e7lar\u0131 ne kadar iyi tahmin edebildi\u011fini anlayabiliriz.<\/span><\/span><\/p>\n Neyse ki, Python’da bir s\u0131n\u0131fland\u0131rma modeli yerle\u015ftirirken, bu \u00fc\u00e7 \u00f6l\u00e7\u00fcm\u00fc olu\u015fturmak i\u00e7in sklearn<\/strong> k\u00fct\u00fcphanesindeki class_report()<\/strong> fonksiyonunu kullanabiliriz.<\/span><\/p>\n A\u015fa\u011f\u0131daki \u00f6rnekte bu fonksiyonun pratikte nas\u0131l kullan\u0131laca\u011f\u0131 g\u00f6sterilmektedir.<\/span><\/p>\n Bu \u00f6rnek i\u00e7in, 1000 farkl\u0131 \u00fcniversite basketbol oyuncusunun NBA’e al\u0131n\u0131p al\u0131nmayaca\u011f\u0131n\u0131 tahmin etmek i\u00e7in puanlar\u0131 ve asistleri kullanan bir lojistik regresyon modeli uygulayaca\u011f\u0131z.<\/span><\/p>\n \u00d6ncelikle Python’da lojistik regresyon ger\u00e7ekle\u015ftirmek i\u00e7in gerekli paketleri i\u00e7e aktaraca\u011f\u0131z:<\/span><\/p>\n Daha sonra 1000 basketbolcunun bilgilerini i\u00e7eren veri \u00e7er\u00e7evesini olu\u015fturaca\u011f\u0131z:<\/span><\/p>\n Not<\/strong> : 0<\/strong> de\u011feri, oyuncunun draft edilmedi\u011fini, 1<\/strong> de\u011feri ise oyuncunun draft edildi\u011fini g\u00f6sterir.<\/span><\/p>\n Daha sonra verilerimizi bir e\u011fitim seti ve bir test seti olarak ay\u0131raca\u011f\u0131z ve lojistik regresyon modeline uyarlayaca\u011f\u0131z:<\/span><\/p>\n Son olarak modelimizin s\u0131n\u0131fland\u0131rma metriklerini yazd\u0131rmak i\u00e7in class_report()<\/strong> fonksiyonunu kullanaca\u011f\u0131z:<\/span><\/p>\n Sonucun nas\u0131l yorumlanaca\u011f\u0131 a\u015fa\u011f\u0131da a\u00e7\u0131klanm\u0131\u015ft\u0131r:<\/span><\/p>\n A\u00e7\u0131klama<\/strong> : Modellerinin draft edilece\u011fini \u00f6ng\u00f6rd\u00fc\u011f\u00fc oyuncular\u0131n yaln\u0131zca %43’\u00fc<\/strong> ger\u00e7ekten \u00f6yleydi.<\/span><\/p>\n Hat\u0131rlatma<\/strong> : Asl\u0131nda se\u00e7ilen t\u00fcm oyuncular aras\u0131nda model, bu sonucu yaln\u0131zca %36’s\u0131<\/strong> i\u00e7in do\u011fru bir \u015fekilde tahmin etti.<\/span><\/p>\n F1 Puan\u0131<\/strong> : Bu de\u011fer \u015fu \u015fekilde hesaplan\u0131r:<\/span><\/p>\n Bu de\u011fer 1’e \u00e7ok yak\u0131n olmad\u0131\u011f\u0131 i\u00e7in bu bize modelin oyuncular\u0131n draft edilip edilmeyece\u011fini tahmin etme konusunda zay\u0131f oldu\u011funu g\u00f6steriyor.<\/span><\/p>\n Destek<\/strong> : Bu de\u011ferler bize basit\u00e7e test veri setindeki her s\u0131n\u0131fa ka\u00e7 oyuncunun ait oldu\u011funu s\u00f6yler. Test veri setindeki oyuncular\u0131n 160’\u0131n\u0131n<\/strong> draft edilmedi\u011fini, 140’\u0131n\u0131n<\/strong> ise draft edilmedi\u011fini g\u00f6rebiliyoruz.<\/span><\/p>\n Not<\/strong> : class_report()<\/strong> i\u015flevine ili\u015fkin t\u00fcm belgeleri burada<\/a> bulabilirsiniz.<\/span><\/p>\n A\u015fa\u011f\u0131daki e\u011fitimler Python’da s\u0131n\u0131fland\u0131rma modellerinin kullan\u0131m\u0131na ili\u015fkin ek bilgiler sa\u011flar:<\/span><\/p>\n Python’da Lojistik Regresyon Nas\u0131l Ger\u00e7ekle\u015ftirilir<\/a> Makine \u00f6\u011freniminde s\u0131n\u0131fland\u0131rma modellerini kulland\u0131\u011f\u0131m\u0131zda modelin kalitesini de\u011ferlendirmek i\u00e7in \u00fc\u00e7 ortak \u00f6l\u00e7\u00fcm kullan\u0131r\u0131z: 1. Do\u011fruluk : Toplam pozitif tahminlere k\u0131yasla do\u011fru pozitif tahminlerin y\u00fczdesi. 2. Hat\u0131rlama : Toplam fiili pozitif tahminlere k\u0131yasla do\u011fru pozitif tahminlerin y\u00fczdesi. 3. F1 Puan\u0131 : Hassasiyet ve hat\u0131rlaman\u0131n a\u011f\u0131rl\u0131kl\u0131 harmonik ortalamas\u0131. Model 1’e ne kadar yak\u0131nsa model o kadar iyidir. […]<\/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-3120","post","type-post","status-publish","format-standard","hentry","category-rehber"],"yoast_head":"\n\n
\u00d6rnek: Sklearn’de s\u0131n\u0131fland\u0131rma raporu nas\u0131l kullan\u0131l\u0131r?<\/strong><\/span><\/h3>\n
import<\/span> pandas as<\/span> pd\nimport<\/span> numpy as<\/span> np\nfrom<\/span> sklearn. model_selection<\/span> import<\/span> train_test_split\nfrom<\/span> sklearn. linear_model<\/span> import<\/span> LogisticRegression\nfrom<\/span> sklearn. metrics<\/span> import<\/span> classification_report<\/strong>\n<\/pre>\n #make this example reproducible\n<\/span>n.p. random<\/span> . seeds<\/span> (1)\n\n#createDataFrame\n<\/span>df = pd. DataFrame<\/span> ({' points<\/span> ': np. random<\/span> . randint<\/span> (30, size=1000),\n ' assists<\/span> ': np. random<\/span> . randint<\/span> (12, size=1000),\n ' drafted<\/span> ': np. random<\/span> . randint<\/span> (2, size=1000)})\n\n#view DataFrame\n<\/span>df. head<\/span> ()\n\n\tpoints assists drafted\n0 5 1 1\n1 11 8 0\n2 12 4 1\n3 8 7 0\n4 9 0 0\n<\/strong><\/span><\/pre>\n #define the predictor variables and the response variable\n<\/span>X = df[[' points<\/span> ', ' assists<\/span> ']]\ny = df[' drafted<\/span> ']\n\n#split the dataset into training (70%) and testing (30%) sets\n<\/span>X_train,X_test,y_train,y_test = train_test_split<\/span> (X,y,test_size=0.3,random_state=0) \n\n<\/strong>#instantiate the model\n<\/span>logistic_regression = LogisticRegression()\n\n#fit the model using the training data\n<\/span>logistic_regression. fit<\/span> (X_train,y_train)\n\n#use model to make predictions on test data\n<\/span>y_pred = logistic_regression. predict<\/span> (X_test)<\/strong><\/pre>\n #print classification report for model\nprint<\/span> (classification_report(y_test, y_pred))\n\n precision recall f1-score support\n\n 0 0.51 0.58 0.54 160\n 1 0.43 0.36 0.40 140\n\n accuracy 0.48 300\n macro avg 0.47 0.47 0.47 300\nweighted avg 0.47 0.48 0.47 300\n<\/span><\/span><\/strong><\/pre>\n\n
Ek kaynaklar<\/strong><\/span><\/h3>\n
Python’da Kar\u0131\u015f\u0131kl\u0131k Matrisi Nas\u0131l Olu\u015fturulur<\/a>
Python’da dengeli hassasiyet nas\u0131l hesaplan\u0131r<\/a><\/p>\n","protected":false},"excerpt":{"rendered":"