{"id":476484,"date":"2023-08-09T07:29:55","date_gmt":"2023-08-09T07:29:55","guid":{"rendered":""},"modified":"2023-09-05T11:12:51","modified_gmt":"2023-09-05T11:12:51","slug":"cross-validation","status":"publish","type":"wiki","link":"https:\/\/oneproxy.pro\/tr\/wiki\/cross-validation\/","title":{"rendered":"\u00c7apraz do\u011frulama"},"content":{"rendered":"<p>\u00c7apraz Do\u011frulama, makine \u00f6\u011frenimi modellerinin performans\u0131n\u0131 de\u011ferlendirmek ve do\u011frulu\u011funu do\u011frulamak i\u00e7in kullan\u0131lan g\u00fc\u00e7l\u00fc bir istatistiksel tekniktir. Tahmine dayal\u0131 modellerin e\u011fitilmesinde ve test edilmesinde \u00e7ok \u00f6nemli bir rol oynar, fazla uyumun \u00f6nlenmesine yard\u0131mc\u0131 olur ve sa\u011flaml\u0131k sa\u011flar. \u00c7apraz Do\u011frulama, veri k\u00fcmesini e\u011fitim ve test i\u00e7in alt k\u00fcmelere b\u00f6lerek, bir modelin g\u00f6r\u00fcnmeyen verilere genelleme yapma becerisine ili\u015fkin daha ger\u00e7ek\u00e7i bir tahmin sa\u011flar.<\/p>\n<h2>\u00c7apraz Do\u011frulaman\u0131n k\u00f6keninin tarihi ve ondan ilk s\u00f6z.<\/h2>\n<p>\u00c7apraz Do\u011frulaman\u0131n k\u00f6kleri istatistik alan\u0131ndad\u0131r ve tarihi 20. y\u00fczy\u0131l\u0131n ortalar\u0131na kadar uzan\u0131r. \u00c7apraz Do\u011frulaman\u0131n ilk s\u00f6z\u00fc, Arthur Bowker ve S. James&#039;in 1949&#039;daki \u00e7al\u0131\u015fmalar\u0131na kadar uzanabilir; burada istatistiksel modellerde \u00f6nyarg\u0131 ve varyans\u0131 tahmin etmek i\u00e7in &quot;jackknife&quot; ad\u0131 verilen bir y\u00f6ntem tan\u0131mlad\u0131lar. Daha sonra, 1968&#039;de John W. Tukey, \u00e7ak\u0131 y\u00f6nteminin bir genellemesi olarak &quot;\u00e7ak\u0131&quot; terimini tan\u0131tt\u0131. Verileri do\u011frulama i\u00e7in alt k\u00fcmelere b\u00f6lme fikri zamanla geli\u015ftirildi ve \u00e7e\u015fitli \u00c7apraz Do\u011frulama tekniklerinin geli\u015ftirilmesine yol a\u00e7t\u0131.<\/p>\n<h2>\u00c7apraz Do\u011frulama hakk\u0131nda detayl\u0131 bilgi. \u00c7apraz Do\u011frulama konusunu geni\u015fletiyoruz.<\/h2>\n<p>\u00c7apraz Do\u011frulama, veri k\u00fcmesini genellikle &quot;katlamalar&quot; olarak adland\u0131r\u0131lan birden fazla alt k\u00fcmeye b\u00f6lerek \u00e7al\u0131\u015f\u0131r. S\u00fcre\u00e7, modelin verinin bir k\u0131sm\u0131 (e\u011fitim seti) \u00fczerinde yinelemeli olarak e\u011fitilmesini ve geri kalan veriler (test seti) \u00fczerindeki performans\u0131n\u0131n de\u011ferlendirilmesini i\u00e7erir. Bu yineleme, her katlama hem e\u011fitim hem de test seti olarak kullan\u0131l\u0131ncaya ve sonu\u00e7lar\u0131n ortalamas\u0131 al\u0131narak nihai bir performans \u00f6l\u00e7\u00fcs\u00fc elde edilene kadar devam eder.<\/p>\n<p>\u00c7apraz Do\u011frulaman\u0131n temel amac\u0131, bir modelin genelleme yetene\u011fini de\u011ferlendirmek ve a\u015f\u0131r\u0131 uyum veya yetersiz uyum gibi potansiyel sorunlar\u0131 belirlemektir. Hiperparametrelerin ayarlanmas\u0131na ve belirli bir sorun i\u00e7in en iyi modelin se\u00e7ilmesine yard\u0131mc\u0131 olur, b\u00f6ylece modelin g\u00f6r\u00fcnmeyen veriler \u00fczerindeki performans\u0131n\u0131 art\u0131r\u0131r.<\/p>\n<h2>\u00c7apraz Do\u011frulaman\u0131n i\u00e7 yap\u0131s\u0131. \u00c7apraz Do\u011frulama nas\u0131l \u00e7al\u0131\u015f\u0131r?<\/h2>\n<p>\u00c7apraz Do\u011frulaman\u0131n i\u00e7 yap\u0131s\u0131 birka\u00e7 ad\u0131mda a\u00e7\u0131klanabilir:<\/p>\n<ol>\n<li>\n<p><strong>Veri B\u00f6lme<\/strong>: Ba\u015flang\u0131\u00e7 veri k\u00fcmesi rastgele k adet e\u015fit boyutlu alt k\u00fcmeye veya katlamaya b\u00f6l\u00fcn\u00fcr.<\/p>\n<\/li>\n<li>\n<p><strong>Model E\u011fitimi ve De\u011ferlendirme<\/strong>: Model k-1 k\u0131vr\u0131m\u0131 \u00fczerinde e\u011fitilir ve kalan k\u0131vr\u0131m \u00fczerinde de\u011ferlendirilir. Bu i\u015flem, her seferinde test seti olarak farkl\u0131 bir katlama kullan\u0131larak k kez tekrarlan\u0131r.<\/p>\n<\/li>\n<li>\n<p><strong>Performans Metri\u011fi<\/strong>: Modelin performans\u0131 do\u011fruluk, kesinlik, geri \u00e7a\u011f\u0131rma, F1 puan\u0131 veya di\u011ferleri gibi \u00f6nceden tan\u0131mlanm\u0131\u015f bir \u00f6l\u00e7\u00fcm kullan\u0131larak \u00f6l\u00e7\u00fcl\u00fcr.<\/p>\n<\/li>\n<li>\n<p><strong>Ortalama Performans<\/strong>: Her yinelemeden elde edilen performans \u00f6l\u00e7\u00fcmlerinin ortalamas\u0131 al\u0131narak tek bir genel performans de\u011feri elde edilir.<\/p>\n<\/li>\n<\/ol>\n<h2>\u00c7apraz Do\u011frulaman\u0131n temel \u00f6zelliklerinin analizi.<\/h2>\n<p>\u00c7apraz Do\u011frulama, onu makine \u00f6\u011frenimi s\u00fcrecinde \u00f6nemli bir ara\u00e7 haline getiren \u00e7e\u015fitli temel \u00f6zellikler sunar:<\/p>\n<ol>\n<li>\n<p><strong>\u00d6nyarg\u0131 Azaltma<\/strong>: \u00c7apraz Do\u011frulama, test i\u00e7in birden fazla alt k\u00fcme kullanarak \u00f6nyarg\u0131y\u0131 azalt\u0131r ve modelin performans\u0131na ili\u015fkin daha do\u011fru bir tahmin sa\u011flar.<\/p>\n<\/li>\n<li>\n<p><strong>Optimum Parametre Ayar\u0131<\/strong>: Bir model i\u00e7in en uygun hiperparametrelerin bulunmas\u0131na yard\u0131mc\u0131 olarak tahmin yetene\u011fini art\u0131r\u0131r.<\/p>\n<\/li>\n<li>\n<p><strong>Sa\u011flaml\u0131k<\/strong>: \u00c7apraz Do\u011frulama, verilerin \u00e7e\u015fitli alt k\u00fcmelerinde tutarl\u0131 bir \u015fekilde iyi performans g\u00f6steren modellerin belirlenmesine yard\u0131mc\u0131 olarak onlar\u0131 daha sa\u011flam hale getirir.<\/p>\n<\/li>\n<li>\n<p><strong>Veri Verimlili\u011fi<\/strong>: Her veri noktas\u0131 hem e\u011fitim hem de do\u011frulama i\u00e7in kullan\u0131ld\u0131\u011f\u0131ndan, mevcut verilerin kullan\u0131m\u0131n\u0131 en \u00fcst d\u00fczeye \u00e7\u0131kar\u0131r.<\/p>\n<\/li>\n<\/ol>\n<h2>\u00c7apraz Do\u011frulama T\u00fcrleri<\/h2>\n<p>Her birinin g\u00fc\u00e7l\u00fc y\u00f6nleri ve uygulamalar\u0131 olan \u00e7e\u015fitli \u00c7apraz Do\u011frulama teknikleri vard\u0131r. \u0130\u015fte yayg\u0131n olarak kullan\u0131lanlardan baz\u0131lar\u0131:<\/p>\n<ol>\n<li>\n<p><strong>K Katlamal\u0131 \u00c7apraz Do\u011frulama<\/strong>: Veri k\u00fcmesi k adet alt k\u00fcmeye b\u00f6l\u00fcn\u00fcr ve model, her yinelemede test k\u00fcmesi olarak farkl\u0131 bir katlama kullan\u0131larak k kez e\u011fitilir ve de\u011ferlendirilir.<\/p>\n<\/li>\n<li>\n<p><strong>Bir \u00c7\u0131k\u0131\u015fl\u0131 \u00c7apraz Do\u011frulama (LOOCV)<\/strong>: K&#039;n\u0131n veri k\u00fcmesindeki veri noktalar\u0131n\u0131n say\u0131s\u0131na e\u015fit oldu\u011fu \u00f6zel bir K-Katlama CV durumu. Her yinelemede yaln\u0131zca bir veri noktas\u0131 test i\u00e7in kullan\u0131l\u0131rken geri kalan\u0131 e\u011fitim i\u00e7in kullan\u0131l\u0131r.<\/p>\n<\/li>\n<li>\n<p><strong>Katmanl\u0131 K-Katlamal\u0131 \u00c7apraz Do\u011frulama<\/strong>: Her katlaman\u0131n orijinal veri k\u00fcmesiyle ayn\u0131 s\u0131n\u0131f da\u011f\u0131l\u0131m\u0131n\u0131 korumas\u0131n\u0131 sa\u011flar; bu, \u00f6zellikle dengesiz veri k\u00fcmeleriyle u\u011fra\u015f\u0131rken faydal\u0131d\u0131r.<\/p>\n<\/li>\n<li>\n<p><strong>Zaman Serisi \u00c7apraz Do\u011frulamas\u0131<\/strong>: E\u011fitim ve test setlerinin kronolojik s\u0131raya g\u00f6re b\u00f6l\u00fcnd\u00fc\u011f\u00fc zaman serisi verileri i\u00e7in \u00f6zel olarak tasarlanm\u0131\u015ft\u0131r.<\/p>\n<\/li>\n<\/ol>\n<h2>\u00c7apraz Do\u011frulamay\u0131 kullanma yollar\u0131, kullan\u0131ma ili\u015fkin sorunlar ve \u00e7\u00f6z\u00fcmleri.<\/h2>\n<p>\u00c7apraz Do\u011frulama a\u015fa\u011f\u0131dakiler gibi \u00e7e\u015fitli senaryolarda yayg\u0131n olarak kullan\u0131l\u0131r:<\/p>\n<ol>\n<li>\n<p><strong>Model Se\u00e7imi<\/strong>: Farkl\u0131 modelleri kar\u015f\u0131la\u015ft\u0131rmaya ve performanslar\u0131na g\u00f6re en iyi olan\u0131 se\u00e7meye yard\u0131mc\u0131 olur.<\/p>\n<\/li>\n<li>\n<p><strong>Hiperparametre Ayar\u0131<\/strong>: \u00c7apraz Do\u011frulama, bir modelin performans\u0131n\u0131 \u00f6nemli \u00f6l\u00e7\u00fcde etkileyen hiperparametrelerin optimum de\u011ferlerinin bulunmas\u0131na yard\u0131mc\u0131 olur.<\/p>\n<\/li>\n<li>\n<p><strong>\u00d6znitelik Se\u00e7imi<\/strong>: \u00c7apraz Do\u011frulama, modelleri farkl\u0131 \u00f6zellik alt k\u00fcmeleriyle kar\u015f\u0131la\u015ft\u0131rarak en alakal\u0131 \u00f6zelliklerin belirlenmesine yard\u0131mc\u0131 olur.<\/p>\n<\/li>\n<\/ol>\n<p>Ancak \u00c7apraz Do\u011frulamayla ilgili baz\u0131 yayg\u0131n sorunlar vard\u0131r:<\/p>\n<ol>\n<li>\n<p><strong>Veri s\u0131z\u0131nt\u0131s\u0131<\/strong>: \u00c7apraz Do\u011frulamadan \u00f6nce \u00f6l\u00e7eklendirme veya \u00f6zellik m\u00fchendisli\u011fi gibi veri \u00f6n i\u015fleme ad\u0131mlar\u0131 uygulan\u0131rsa, test k\u00fcmesindeki bilgiler yanl\u0131\u015fl\u0131kla e\u011fitim s\u00fcrecine s\u0131zarak tarafl\u0131 sonu\u00e7lara yol a\u00e7abilir.<\/p>\n<\/li>\n<li>\n<p><strong>Hesaplamal\u0131 Maliyet<\/strong>: \u00c7apraz Do\u011frulama, \u00f6zellikle b\u00fcy\u00fck veri k\u00fcmeleri veya karma\u015f\u0131k modellerle u\u011fra\u015f\u0131rken hesaplama a\u00e7\u0131s\u0131ndan pahal\u0131 olabilir.<\/p>\n<\/li>\n<\/ol>\n<p>Bu sorunlar\u0131n \u00fcstesinden gelmek i\u00e7in ara\u015ft\u0131rmac\u0131lar ve uygulay\u0131c\u0131lar genellikle uygun veri \u00f6n i\u015fleme, paralelle\u015ftirme ve \u00c7apraz Do\u011frulama d\u00f6ng\u00fcs\u00fc i\u00e7inde \u00f6zellik se\u00e7imi gibi teknikleri kullan\u0131r.<\/p>\n<h2>Ana \u00f6zellikler ve benzer terimlerle di\u011fer kar\u015f\u0131la\u015ft\u0131rmalar tablo ve liste \u015feklinde.<\/h2>\n<table>\n<thead>\n<tr>\n<th>\u00d6zellikler<\/th>\n<th>\u00c7apraz do\u011frulama<\/th>\n<th>\u00d6ny\u00fckleme<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td>Ama\u00e7<\/td>\n<td>Model de\u011ferlendirmesi<\/td>\n<td>Parametre tahmini<\/td>\n<\/tr>\n<tr>\n<td>Veri B\u00f6lme<\/td>\n<td>\u00c7oklu katlama<\/td>\n<td>Rasgele \u00f6rnekleme<\/td>\n<\/tr>\n<tr>\n<td>Yinelemeler<\/td>\n<td>k kere<\/td>\n<td>Yeniden \u00f6rnekleme<\/td>\n<\/tr>\n<tr>\n<td>Performans Tahmini<\/td>\n<td>Ortalama<\/td>\n<td>Y\u00fczdelikler<\/td>\n<\/tr>\n<tr>\n<td>Kullan\u0131m Durumlar\u0131<\/td>\n<td>Model se\u00e7imi<\/td>\n<td>Belirsizlik tahmini<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<p><strong>Bootstrapping ile Kar\u015f\u0131la\u015ft\u0131rma<\/strong>:<\/p>\n<ul>\n<li>\u00c7apraz Do\u011frulama \u00f6ncelikle model de\u011ferlendirmesi i\u00e7in kullan\u0131l\u0131rken Bootstrap daha \u00e7ok parametre tahmini ve belirsizlik \u00f6l\u00e7\u00fcm\u00fcne odaklan\u0131r.<\/li>\n<li>\u00c7apraz Do\u011frulama, verileri birden fazla b\u00f6l\u00fcme ay\u0131rmay\u0131 i\u00e7erirken Bootstrap, verileri de\u011fi\u015ftirerek rastgele \u00f6rnekler.<\/li>\n<\/ul>\n<h2>\u00c7apraz Do\u011frulama ile ilgili gelece\u011fin perspektifleri ve teknolojileri.<\/h2>\n<p>\u00c7apraz Do\u011frulaman\u0131n gelece\u011fi, geli\u015fmi\u015f makine \u00f6\u011frenimi teknikleri ve teknolojileriyle entegrasyonunda yatmaktad\u0131r:<\/p>\n<ol>\n<li>\n<p><strong>Derin \u00d6\u011frenme Entegrasyonu<\/strong>: \u00c7apraz Do\u011frulamay\u0131 derin \u00f6\u011frenme yakla\u015f\u0131mlar\u0131yla birle\u015ftirmek, karma\u015f\u0131k sinir a\u011flar\u0131 i\u00e7in model de\u011ferlendirmesini ve hiper parametre ayarlamas\u0131n\u0131 geli\u015ftirecektir.<\/p>\n<\/li>\n<li>\n<p><strong>Otomatik ML<\/strong>: Otomatik Makine \u00d6\u011frenimi (AutoML) platformlar\u0131, makine \u00f6\u011frenimi modellerinin se\u00e7imini ve yap\u0131land\u0131rmas\u0131n\u0131 optimize etmek i\u00e7in \u00c7apraz Do\u011frulamadan yararlanabilir.<\/p>\n<\/li>\n<li>\n<p><strong>Paralelle\u015ftirme<\/strong>: Paralel bilgi i\u015flem ve da\u011f\u0131t\u0131lm\u0131\u015f sistemlerden yararlanmak, \u00c7apraz Do\u011frulamay\u0131 b\u00fcy\u00fck veri k\u00fcmeleri i\u00e7in daha \u00f6l\u00e7eklenebilir ve verimli hale getirecektir.<\/p>\n<\/li>\n<\/ol>\n<h2>Proxy sunucular\u0131 nas\u0131l kullan\u0131labilir veya \u00c7apraz Do\u011frulama ile nas\u0131l ili\u015fkilendirilebilir?<\/h2>\n<p>Proxy sunucular\u0131 internetle ilgili \u00e7e\u015fitli uygulamalarda \u00f6nemli bir rol oynar ve \u00c7apraz Do\u011frulama ile a\u015fa\u011f\u0131daki yollarla ili\u015fkilendirilebilirler:<\/p>\n<ol>\n<li>\n<p><strong>Veri toplama<\/strong>: Proxy sunucular\u0131, tarafs\u0131z \u00c7apraz Do\u011frulama sonu\u00e7lar\u0131 i\u00e7in gerekli olan, \u00e7e\u015fitli co\u011frafi konumlardan \u00e7e\u015fitli veri k\u00fcmelerini toplamak i\u00e7in kullan\u0131labilir.<\/p>\n<\/li>\n<li>\n<p><strong>G\u00fcvenlik ve Gizlilik<\/strong>: Hassas verilerle u\u011fra\u015f\u0131rken, proxy sunucular \u00c7apraz Do\u011frulama s\u0131ras\u0131nda kullan\u0131c\u0131 bilgilerinin anonimle\u015ftirilmesine yard\u0131mc\u0131 olarak veri gizlili\u011fini ve g\u00fcvenli\u011fini sa\u011flayabilir.<\/p>\n<\/li>\n<li>\n<p><strong>Y\u00fck dengeleme<\/strong>: Da\u011f\u0131t\u0131lm\u0131\u015f \u00c7apraz Do\u011frulama kurulumlar\u0131nda, proxy sunucular farkl\u0131 d\u00fc\u011f\u00fcmler aras\u0131nda y\u00fck dengelemeye yard\u0131mc\u0131 olarak hesaplama verimlili\u011fini art\u0131rabilir.<\/p>\n<\/li>\n<\/ol>\n<h2>\u0130lgili Ba\u011flant\u0131lar<\/h2>\n<p>\u00c7apraz Do\u011frulama hakk\u0131nda daha fazla bilgi i\u00e7in a\u015fa\u011f\u0131daki kaynaklara ba\u015fvurabilirsiniz:<\/p>\n<ol>\n<li><a href=\"https:\/\/scikit-learn.org\/stable\/modules\/cross_validation.html\" target=\"_new\" rel=\"noopener nofollow\">Scikit-learn \u00c7apraz Do\u011frulama Belgeleri<\/a><\/li>\n<li><a href=\"https:\/\/towardsdatascience.com\/a-gentle-introduction-to-cross-validation-209a89d69c55\" target=\"_new\" rel=\"noopener nofollow\">Veri Bilimine Do\u011fru \u2013 \u00c7apraz Do\u011frulamaya Nazik Bir Giri\u015f<\/a><\/li>\n<li><a href=\"https:\/\/en.wikipedia.org\/wiki\/Cross-validation\" target=\"_new\" rel=\"noopener nofollow\">Vikipedi \u2013 \u00c7apraz Do\u011frulama<\/a><\/li>\n<\/ol>","protected":false},"featured_media":468046,"menu_order":0,"template":"","meta":{"_acf_changed":false,"content-type":"","inline_featured_image":false,"footnotes":""},"class_list":["post-476484","wiki","type-wiki","status-publish","has-post-thumbnail","hentry"],"acf":{"faq_title":"Frequently Asked Questions about <mark>Cross-Validation: Understanding the Power of Validation Techniques<\/mark>","faq_items":[{"question":"What is Cross-Validation, and why is it important in machine learning?","answer":"<p>Cross-Validation is a statistical technique used to assess the performance of machine learning models by partitioning the dataset into subsets for training and testing. It helps to avoid overfitting and ensures the model's ability to generalize to new data. By providing a more realistic estimation of model performance, Cross-Validation plays a vital role in selecting the best model and tuning hyperparameters.<\/p>"},{"question":"How does Cross-Validation work?","answer":"<p>Cross-Validation involves dividing the data into k subsets or folds. The model is trained on k-1 folds and evaluated on the remaining one, iterating this process k times with each fold serving as the test set once. The final performance metric is an average of the metrics obtained in each iteration.<\/p>"},{"question":"What are the different types of Cross-Validation?","answer":"<p>Some common types of Cross-Validation include K-Fold Cross-Validation, Leave-One-Out Cross-Validation (LOOCV), Stratified K-Fold Cross-Validation, and Time Series Cross-Validation. Each type has specific use cases and advantages.<\/p>"},{"question":"What are the key benefits of using Cross-Validation?","answer":"<p>Cross-Validation offers several benefits, including bias reduction, optimal parameter tuning, robustness, and maximum data efficiency. It helps in identifying models that perform consistently well and improves the model's reliability.<\/p>"},{"question":"How can Cross-Validation be used in machine learning?","answer":"<p>Cross-Validation is used for various purposes, such as model selection, hyperparameter tuning, and feature selection. It provides valuable insights into a model's performance and aids in making better decisions during the model development process.<\/p>"},{"question":"What are the potential problems related to Cross-Validation and their solutions?","answer":"<p>Some common issues with Cross-Validation include data leakage and computational cost. To address these problems, practitioners can apply proper data preprocessing techniques and leverage parallelization for efficient execution.<\/p>"},{"question":"How does Cross-Validation compare to Bootstrap?","answer":"<p>Cross-Validation is primarily used for model evaluation, while Bootstrap focuses on parameter estimation and uncertainty quantification. Cross-Validation involves multiple folds, while Bootstrap uses random sampling with replacement.<\/p>"},{"question":"What does the future hold for Cross-Validation in the machine learning landscape?","answer":"<p>The future of Cross-Validation involves integration with advanced machine learning techniques, like deep learning and AutoML. Leveraging parallel computing and distributed systems will make Cross-Validation more scalable and efficient.<\/p>"},{"question":"How do proxy servers relate to Cross-Validation?","answer":"<p>Proxy servers can be associated with Cross-Validation in data collection, security, and load balancing. They help in collecting diverse datasets, ensuring data privacy, and optimizing distributed Cross-Validation setups.<\/p>"}]},"_links":{"self":[{"href":"https:\/\/oneproxy.pro\/tr\/wp-json\/wp\/v2\/wiki\/476484","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/oneproxy.pro\/tr\/wp-json\/wp\/v2\/wiki"}],"about":[{"href":"https:\/\/oneproxy.pro\/tr\/wp-json\/wp\/v2\/types\/wiki"}],"version-history":[{"count":0,"href":"https:\/\/oneproxy.pro\/tr\/wp-json\/wp\/v2\/wiki\/476484\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/oneproxy.pro\/tr\/wp-json\/wp\/v2\/media\/468046"}],"wp:attachment":[{"href":"https:\/\/oneproxy.pro\/tr\/wp-json\/wp\/v2\/media?parent=476484"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}