{"id":6232,"date":"2020-01-16T11:34:29","date_gmt":"2020-01-16T04:34:29","guid":{"rendered":"http:\/\/gcloudvn.wam.vn\/phan-tich-du-lieu-lua-chon-tinh-nang-cho-cac-mo-hinh-ml-tot-hon\/"},"modified":"2023-04-27T11:19:30","modified_gmt":"2023-04-27T04:19:30","slug":"phan-tich-du-lieu-lua-chon-tinh-nang-cho-cac-mo-hinh-ml-tot-hon","status":"publish","type":"post","link":"https:\/\/gcloudvn.com\/en\/kienthuc\/phan-tich-du-lieu-lua-chon-tinh-nang-cho-cac-mo-hinh-ml-tot-hon\/","title":{"rendered":"Data analysis, feature selection for better ML models"},"content":{"rendered":"<p style=\"text-align: justify;\"><span style=\"font-weight: 400;\">Khi b\u1ea1n b\u1eaft \u0111\u1ea7u v\u1edbi m\u1ed9t d\u1ef1 \u00e1n m\u00e1y h\u1ecdc (ML), m\u1ed9t nguy\u00ean t\u1eafc quan tr\u1ecdng c\u1ea7n ghi nh\u1edb l\u00e0 d\u1eef li\u1ec7u l\u00e0 t\u1ea5t c\u1ea3. Ng\u01b0\u1eddi ta th\u01b0\u1eddng n\u00f3i r\u1eb1ng n\u1ebfu ML l\u00e0 \u0111\u1ed9ng c\u01a1 t\u00ean l\u1eeda, th\u00ec nhi\u00ean li\u1ec7u l\u00e0 d\u1eef li\u1ec7u (ch\u1ea5t l\u01b0\u1ee3ng cao) \u0111\u01b0\u1ee3c cung c\u1ea5p cho thu\u1eadt to\u00e1n ML. Tuy nhi\u00ean, t\u00ecm ra s\u1ef1 th\u1eadt v\u00e0 c\u00e1i nh\u00ecn s\u00e2u s\u1eafc t\u1eeb m\u1ed9t \u0111\u1ed1ng d\u1eef li\u1ec7u c\u00f3 th\u1ec3 l\u00e0 m\u1ed9t c\u00f4ng vi\u1ec7c ph\u1ee9c t\u1ea1p v\u00e0 d\u1ec5 b\u1ecb l\u1ed7i. \u0110\u1ec3 c\u00f3 m\u1ed9t kh\u1edfi \u0111\u1ea7u v\u1eefng ch\u1eafc cho d\u1ef1 \u00e1n ML c\u1ee7a b\u1ea1n, vi\u1ec7c ph\u00e2n t\u00edch d\u1eef li\u1ec7u tr\u01b0\u1edbc l\u00e0 h\u1ebft s\u1ee9c quan tr\u1ecdng, m\u1ed9t trong c\u00e1c b\u01b0\u1edbc l\u00e0 m\u00f4 t\u1ea3 d\u1eef li\u1ec7u b\u1eb1ng c\u00e1c k\u1ef9 thu\u1eadt th\u1ed1ng k\u00ea v\u00e0 tr\u1ef1c quan h\u00f3a \u0111\u1ec3 \u0111\u01b0a c\u00e1c kh\u00eda c\u1ea1nh quan tr\u1ecdng c\u1ee7a d\u1eef li\u1ec7u \u0111\u00f3 v\u00e0o tr\u1ecdng t\u00e2m \u0111\u1ec3 ph\u00e2n t\u00edch th\u00eam. Trong qu\u00e1 tr\u00ecnh \u0111\u00f3, \u0111i\u1ec1u quan tr\u1ecdng l\u00e0 b\u1ea1n ph\u1ea3i hi\u1ec3u th\u1eadt k\u0129 v\u1ec1:\u00a0<\/span><\/p>\n<ul style=\"text-align: justify;\">\n<li style=\"font-weight: 400;\"><b>Data properties: <\/b><span style=\"font-weight: 400;\">such as schema and statistical properties;<\/span><\/li>\n<li style=\"font-weight: 400;\"><b>Quality of data:<\/b><span style=\"font-weight: 400;\"> like missing values and inconsistent data types;<\/span><\/li>\n<li style=\"font-weight: 400;\"><b>The predictive power of data<\/b><span style=\"font-weight: 400;\">: such as the relationship of features to the target.<\/span><\/li>\n<\/ul>\n<p style=\"text-align: justify;\"><span style=\"font-weight: 400;\">This process is the foundation for subsequent engineering and feature selection steps, and it provides a solid foundation for building better ML models.\u00a0<\/span><\/p>\n<p style=\"text-align: justify;\"><span style=\"font-weight: 400;\">There are many different approaches to conducting exploratory data analysis (EDA) out there, so it can be difficult to know which analysis to do and how to do it right. To reinforce recommendations for conducting EDA, data cleaning, and appropriate feature selection in ML projects, Google summarizes and provides brief guidance from both visual (visual) perspectives. ) and strict (statistical). Based on the results of the analysis, you can then define feature selections and corresponding technical recommendations. You can also see more tutorials by <\/span><a href=\"http:\/\/services.google.com\/fh\/files\/misc\/exploratory_data_analysis_for_feature_selection_in_machine_learning.pdf\" target=\"_blank\" rel=\"nofollow noopener\"><span style=\"font-weight: 400;\">this link<\/span><\/a><span style=\"font-weight: 400;\">.<\/span><\/p>\n<p style=\"text-align: justify;\"><span style=\"font-weight: 400;\">You can also check <\/span><a href=\"https:\/\/github.com\/GoogleCloudPlatform\/professional-services\/tree\/master\/tools\/ml-auto-eda\" target=\"_blank\" rel=\"nofollow noopener\"><span style=\"font-weight: 400;\">Automatic data discovery and feature recommendation engine<\/span><\/a><span style=\"font-weight: 400;\"> that Google has developed to help you automate recommended analysis, regardless of the size of your data, and then generate well-organized reports that present the results.\u00a0<\/span><\/p>\n<p style=\"text-align: justify;\"><span style=\"font-weight: 400;\">EDA, feature selection, and feature engineering are often tied together and are important steps in the ML journey. Given the complexity of data and the business issues that exist today (such as credit scoring in finance and demand forecasting in retail), how the results of an appropriate EDA can affect Your next decision is a big question. In this post, Google will walk you through some of the decisions you&#039;ll make about your data for a particular project, and choose what type of analytics to use, along with visualizations, tools, and processing. feature management.<\/span><\/p>\n<p style=\"text-align: justify;\"><span style=\"font-weight: 400;\">Let&#039;s start exploring the types of analytics you can choose from.<\/span><\/p>\n<div id=\"ez-toc-container\" class=\"ez-toc-v2_0_80 counter-hierarchy ez-toc-counter ez-toc-grey ez-toc-container-direction\">\n<div class=\"ez-toc-title-container\">\n<p class=\"ez-toc-title\" style=\"cursor:inherit\">Table of contents<\/p>\n<span class=\"ez-toc-title-toggle\"><a href=\"#\" class=\"ez-toc-pull-right ez-toc-btn ez-toc-btn-xs ez-toc-btn-default ez-toc-toggle\" aria-label=\"Toggle Table of Content\"><span class=\"ez-toc-js-icon-con\"><span class=\"\"><span class=\"eztoc-hide\" style=\"display:none;\">Toggle<\/span><span class=\"ez-toc-icon-toggle-span\"><svg style=\"fill: #999;color:#999\" xmlns=\"http:\/\/www.w3.org\/2000\/svg\" class=\"list-377408\" width=\"20px\" height=\"20px\" viewbox=\"0 0 24 24\" fill=\"none\"><path d=\"M6 6H4v2h2V6zm14 0H8v2h12V6zM4 11h2v2H4v-2zm16 0H8v2h12v-2zM4 16h2v2H4v-2zm16 0H8v2h12v-2z\" fill=\"currentColor\"><\/path><\/svg><svg style=\"fill: #999;color:#999\" class=\"arrow-unsorted-368013\" xmlns=\"http:\/\/www.w3.org\/2000\/svg\" width=\"10px\" height=\"10px\" viewbox=\"0 0 24 24\" version=\"1.2\" baseprofile=\"tiny\"><path d=\"M18.2 9.3l-6.2-6.3-6.2 6.3c-.2.2-.3.4-.3.7s.1.5.3.7c.2.2.4.3.7.3h11c.3 0 .5-.1.7-.3.2-.2.3-.5.3-.7s-.1-.5-.3-.7zM5.8 14.7l6.2 6.3 6.2-6.3c.2-.2.3-.5.3-.7s-.1-.5-.3-.7c-.2-.2-.4-.3-.7-.3h-11c-.3 0-.5.1-.7.3-.2.2-.3.5-.3.7s.1.5.3.7z\"\/><\/svg><\/span><\/span><\/span><\/a><\/span><\/div>\n<nav><ul class='ez-toc-list ez-toc-list-level-1' ><li class='ez-toc-page-1 ez-toc-heading-level-2'><a class=\"ez-toc-link ez-toc-heading-1\" href=\"https:\/\/gcloudvn.com\/en\/kienthuc\/phan-tich-du-lieu-lua-chon-tinh-nang-cho-cac-mo-hinh-ml-tot-hon\/#Phan_tich_du_lieu_thong_ke\" >Statistical data analysis<\/a><ul class='ez-toc-list-level-3' ><li class='ez-toc-heading-level-3'><a class=\"ez-toc-link ez-toc-heading-2\" href=\"https:\/\/gcloudvn.com\/en\/kienthuc\/phan-tich-du-lieu-lua-chon-tinh-nang-cho-cac-mo-hinh-ml-tot-hon\/#1_Phan_tich_mo_ta_Phan_tich_don_bien\" >1. Descriptive analysis (Univariate analysis)<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-3'><a class=\"ez-toc-link ez-toc-heading-3\" href=\"https:\/\/gcloudvn.com\/en\/kienthuc\/phan-tich-du-lieu-lua-chon-tinh-nang-cho-cac-mo-hinh-ml-tot-hon\/#2_Phan_tich_tuong_quan_phan_tich_bivariate\" >2. Correlation analysis (bivariate analysis)<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-3'><a class=\"ez-toc-link ez-toc-heading-4\" href=\"https:\/\/gcloudvn.com\/en\/kienthuc\/phan-tich-du-lieu-lua-chon-tinh-nang-cho-cac-mo-hinh-ml-tot-hon\/#3_Phan_tich_boi_canh\" >3. Context analysis\u00a0<\/a><\/li><\/ul><\/li><li class='ez-toc-page-1 ez-toc-heading-level-2'><a class=\"ez-toc-link ez-toc-heading-5\" href=\"https:\/\/gcloudvn.com\/en\/kienthuc\/phan-tich-du-lieu-lua-chon-tinh-nang-cho-cac-mo-hinh-ml-tot-hon\/#Lua_chon_tinh_nang_va_ki_thuat\" >Selection of features and techniques<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-2'><a class=\"ez-toc-link ez-toc-heading-6\" href=\"https:\/\/gcloudvn.com\/en\/kienthuc\/phan-tich-du-lieu-lua-chon-tinh-nang-cho-cac-mo-hinh-ml-tot-hon\/#Mot_so_cong_cu_giup_ban_tu_dong_hoa\" >Some tools to help you automate<\/a><\/li><\/ul><\/nav><\/div>\n<h2 style=\"text-align: justify;\"><span class=\"ez-toc-section\" id=\"Phan_tich_du_lieu_thong_ke\"><\/span><span style=\"font-weight: 400;\">Statistical data analysis<\/span><span class=\"ez-toc-section-end\"><\/span><\/h2>\n<p style=\"text-align: justify;\"><span style=\"font-weight: 400;\">With this type of analysis, data exploration can be conducted from three different perspectives: descriptive, correlated, and contextual. Each type introduces additional information about the properties and predictability of the data, helping you make informed decisions based on the results of your analysis.<\/span><\/p>\n<h3 style=\"text-align: justify;\"><span class=\"ez-toc-section\" id=\"1_Phan_tich_mo_ta_Phan_tich_don_bien\"><\/span><span style=\"font-weight: 400;\">1. Descriptive analysis (Univariate analysis)<\/span><span class=\"ez-toc-section-end\"><\/span><\/h3>\n<p style=\"text-align: justify;\"><span style=\"font-weight: 400;\">Descriptive analysis, or univariate analysis, provides an understanding of the characteristics of each attribute of a dataset. It also provides important evidence for preprocessing and selection at a later stage. The following table lists the recommended analysis for common, numerical, categorical, and text attributes.<\/span><\/p>\n<p style=\"text-align: justify;\"><img loading=\"lazy\" decoding=\"async\" class=\"aligncenter wp-image-15339 size-full\" src=\"https:\/\/gcloudvn.com\/wp-content\/uploads\/2020\/01\/GCP3.png\" alt=\"Data analysis, feature selection for better ML models 1\" width=\"512\" height=\"288\" \/><\/p>\n<h3 style=\"text-align: justify;\"><span class=\"ez-toc-section\" id=\"2_Phan_tich_tuong_quan_phan_tich_bivariate\"><\/span><span style=\"font-weight: 400;\">2. Correlation analysis (bivariate analysis)<\/span><span class=\"ez-toc-section-end\"><\/span><\/h3>\n<p style=\"text-align: justify;\"><span style=\"font-weight: 400;\">Correlation analysis (or bivariate analysis) tests the relationship between two properties, say X and Y, and checks whether X and Y are correlated. This analysis can be done from two perspectives to get different possible combinations:<\/span><\/p>\n<ul style=\"text-align: justify;\">\n<li style=\"font-weight: 400;\"><b>Qualitative analysis.<\/b><span style=\"font-weight: 400;\"> This performs the calculation of the descriptive statistics of the dependent categorical\/numeric attributes according to each unique value of the independent categorical attribute. This perspective helps to intuitively understand the relationship between X and Y. Visualization is often used in conjunction with qualitative analysis as a more intuitive way to present results.<\/span><\/li>\n<\/ul>\n<p style=\"text-align: justify;\"><img loading=\"lazy\" decoding=\"async\" class=\"aligncenter wp-image-15340 size-full\" src=\"https:\/\/gcloudvn.com\/wp-content\/uploads\/2020\/01\/GCP4.png\" alt=\"Data analysis, feature selection for better ML models 2\" width=\"512\" height=\"83\" \/><\/p>\n<ul style=\"text-align: justify;\">\n<li style=\"font-weight: 400;\"><b>Quantitative analysis.<\/b><span style=\"font-weight: 400;\"> This is a quantitative test of the relationship between X and Y, based on a hypothesis testing framework. This perspective provides a formal and mathematical method for quantitatively determining the existence and\/or strength of a relationship.<\/span><\/li>\n<\/ul>\n<p style=\"text-align: justify;\"><img loading=\"lazy\" decoding=\"async\" class=\"aligncenter wp-image-15343 size-full\" src=\"https:\/\/gcloudvn.com\/wp-content\/uploads\/2020\/01\/GCP7.png\" alt=\"Data analysis, feature selection for better ML models 3\" width=\"512\" height=\"141\" \/><\/p>\n<h3 style=\"text-align: justify;\"><span class=\"ez-toc-section\" id=\"3_Phan_tich_boi_canh\"><\/span><span style=\"font-weight: 400;\">3. Context analysis\u00a0<\/span><span class=\"ez-toc-section-end\"><\/span><\/h3>\n<p style=\"text-align: justify;\"><span style=\"font-weight: 400;\">Descriptive analysis and correlation analysis are both general enough to be performed on any structured data set, neither of which requires context information. To further understand or profile a given dataset, and to understand more about specific domains, you can use one of two common contextual information-based analytics:\u00a0<\/span><\/p>\n<ul style=\"text-align: justify;\">\n<li style=\"font-weight: 400;\"><b>Time-based analysis<\/b><span style=\"font-weight: 400;\">:In many real-world datasets, a timestamp (or similar time-related attribute) is one of the key pieces of contextual information. Observing and\/or understanding the characteristics of data over time, with varying levels of detail, is essential to understanding the data generation process and ensuring data quality<\/span><\/li>\n<li style=\"font-weight: 400;\"><b>Agent-based analysis<\/b><span style=\"font-weight: 400;\">: As an alternative to time, another common attribute is the unique identifier (ID, such as user ID) of each record. Analyzing the dataset by aggregating along agent size, i.e., histogram of the number of records per agent, can further improve your understanding of the dataset.\u00a0<\/span><\/li>\n<\/ul>\n<p style=\"text-align: justify;\"><b>Example of time analysis<\/b><span style=\"font-weight: 400;\">:<\/span><\/p>\n<p style=\"text-align: justify;\"><span style=\"font-weight: 400;\">The figure below shows the average number of trains per hour that originate and end at a specific location based on a simulated dataset<\/span><\/p>\n<p style=\"text-align: justify;\"><img loading=\"lazy\" decoding=\"async\" class=\"aligncenter wp-image-15342 size-full\" src=\"https:\/\/gcloudvn.com\/wp-content\/uploads\/2020\/01\/GCP6.png\" alt=\"Data analysis, feature selection for ML models better than 5\" width=\"512\" height=\"288\" \/><\/p>\n<p style=\"text-align: justify;\"><span style=\"font-weight: 400;\">From this, we can conclude that peak times are between 8:30 a.m. and 5:30 p.m., which is consistent with the intuition that these are the times when people usually leave home in the morning and return to their homes. back after a working day.<\/span><\/p>\n<h2 style=\"text-align: justify;\"><span class=\"ez-toc-section\" id=\"Lua_chon_tinh_nang_va_ki_thuat\"><\/span><span style=\"font-weight: 400;\">Selection of features and techniques<\/span><span class=\"ez-toc-section-end\"><\/span><\/h2>\n<p style=\"text-align: justify;\"><span style=\"font-weight: 400;\">The ultimate goal of EDA (whether rigorously or through visualization) is to provide insights into the dataset you&#039;re studying. This can inspire your next feature selection, technique, and modeling process.\u00a0<\/span><\/p>\n<p style=\"text-align: justify;\"><span style=\"font-weight: 400;\">Descriptive analysis provides basic statistics of each attribute of the dataset. Those statistics can help you identify the following problems:\u00a0<\/span><\/p>\n<ul style=\"text-align: justify;\">\n<li style=\"font-weight: 400;\"><span style=\"font-weight: 400;\">High percentage is missing<\/span><\/li>\n<li style=\"font-weight: 400;\"><span style=\"font-weight: 400;\">Low variance<\/span><\/li>\n<li style=\"font-weight: 400;\"><span style=\"font-weight: 400;\">Low entropy of categorical properties<\/span><\/li>\n<li style=\"font-weight: 400;\"><span style=\"font-weight: 400;\">Classification target imbalance (class imbalance)<\/span><\/li>\n<li style=\"font-weight: 400;\"><span style=\"font-weight: 400;\">Oblique distribution of numeric attributes<\/span><\/li>\n<li style=\"font-weight: 400;\"><span style=\"font-weight: 400;\">High Cardinality of Categorical Attributes<\/span><\/li>\n<\/ul>\n<p style=\"text-align: justify;\"><span style=\"font-weight: 400;\">Correlation analysis examines the relationship between two attributes. There are two typical action points triggered by correlation analysis in the context of feature selection or feature engineering:<\/span><\/p>\n<ul style=\"text-align: justify;\">\n<li style=\"font-weight: 400;\"><span style=\"font-weight: 400;\">Low correlation between features and goals<\/span><\/li>\n<li style=\"font-weight: 400;\"><span style=\"font-weight: 400;\">High correlation between features<\/span><\/li>\n<\/ul>\n<p style=\"text-align: justify;\"><span style=\"font-weight: 400;\">After you have identified the problems, the next task is to make a sound decision on how to properly mitigate these problems. One such example is for the high percentage of missing values. The problem was identified as a missing attribute in a significant proportion of the data points. The threshold or definition of importance can be set based on domain knowledge. There are two options for handling this, depending on the business scenario:<\/span><\/p>\n<ol style=\"text-align: justify;\">\n<li style=\"font-weight: 400;\"><span style=\"font-weight: 400;\">Assigning a unique value to the missing value records, if the missing value, in certain contexts, actually makes sense. For example, a missing value may indicate that a monitored underlying process is not functioning properly.\u00a0<\/span><\/li>\n<li style=\"font-weight: 400;\"><span style=\"font-weight: 400;\">Discard the feature if values are missing due to misconfiguration, data collection problems or uncontrollable random reasons, and historical data can be restored.\u00a0<\/span><\/li>\n<\/ol>\n<p style=\"text-align: justify;\"><span style=\"font-weight: 400;\">You can <\/span><a href=\"http:\/\/services.google.com\/fh\/files\/misc\/exploratory_data_analysis_for_feature_selection_in_machine_learning.pdf\" target=\"_blank\" rel=\"nofollow noopener\"><span style=\"font-weight: 400;\">check the whitepaper<\/span><\/a><span style=\"font-weight: 400;\"> To learn more about ways to solve the above problems, it is recommended to visualize each analysis and examine the most appropriate tools available.<\/span><\/p>\n<h2 style=\"text-align: justify;\"><span class=\"ez-toc-section\" id=\"Mot_so_cong_cu_giup_ban_tu_dong_hoa\"><\/span><span style=\"font-weight: 400;\">Some tools to help you automate<\/span><span class=\"ez-toc-section-end\"><\/span><\/h2>\n<p style=\"text-align: justify;\"><span style=\"font-weight: 400;\">To further help you speed up data preparation for machine learning, you can use <\/span><span style=\"font-weight: 400;\">Automatic data discovery and feature recommendation engine<\/span><span style=\"font-weight: 400;\"> by Google to automate recommended analysis regardless of the size of the data and generate well-organized reports presenting results and recommendations.<\/span><\/p>\n<p style=\"text-align: justify;\"><span style=\"font-weight: 400;\">Automation EDA tools include:<\/span><\/p>\n<ul style=\"text-align: justify;\">\n<li style=\"font-weight: 400;\"><span style=\"font-weight: 400;\">Descriptive analysis of each attribute in a dataset for numerical, categorical;\u00a0<\/span><\/li>\n<li style=\"font-weight: 400;\"><span style=\"font-weight: 400;\">Correlation analysis of two attributes (number versus number, number vs class, and class vs class) through qualitative and\/or quantitative analysis.<\/span><\/li>\n<\/ul>\n<p style=\"text-align: justify;\"><span style=\"font-weight: 400;\">Based on the EDA performed, the tool makes feature recommendations and generates a summary report, which looks something like this:<\/span><\/p>\n<p style=\"text-align: justify;\"><img loading=\"lazy\" decoding=\"async\" class=\"aligncenter wp-image-15341 size-full\" src=\"https:\/\/gcloudvn.com\/wp-content\/uploads\/2020\/01\/GCP5.png\" alt=\"Data analysis, feature selection for better ML models 6\" width=\"512\" height=\"447\" \/><\/p>\n<p style=\"text-align: right;\"><strong>Source: <a href=\"https:\/\/gcloudvn.com\/en\/\">Gimasys<\/a><\/strong><\/p>","protected":false},"excerpt":{"rendered":"<p>When you start with a machine learning (ML) project, an important principle to keep in mind is that data is everything. It is often said that if the ML is a rocket engine, then the fuel\u2026<\/p>","protected":false},"author":1,"featured_media":6233,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"inline_featured_image":false,"footnotes":""},"categories":[1],"tags":[],"class_list":["post-6232","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-kienthuc","entry","has-media"],"_links":{"self":[{"href":"https:\/\/gcloudvn.com\/en\/wp-json\/wp\/v2\/posts\/6232","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/gcloudvn.com\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/gcloudvn.com\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/gcloudvn.com\/en\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/gcloudvn.com\/en\/wp-json\/wp\/v2\/comments?post=6232"}],"version-history":[{"count":0,"href":"https:\/\/gcloudvn.com\/en\/wp-json\/wp\/v2\/posts\/6232\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/gcloudvn.com\/en\/wp-json\/wp\/v2\/media\/6233"}],"wp:attachment":[{"href":"https:\/\/gcloudvn.com\/en\/wp-json\/wp\/v2\/media?parent=6232"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/gcloudvn.com\/en\/wp-json\/wp\/v2\/categories?post=6232"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/gcloudvn.com\/en\/wp-json\/wp\/v2\/tags?post=6232"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}