Mantenha tudo organizado com as coleções
Salve e categorize o conteúdo com base nas suas preferências.
Visão geral do pré-processamento de atributos
O pré-processamento de atributos é uma das etapas mais importantes do ciclo de vida de machine learning. Ela consiste em criar atributos e limpar os dados de treinamento. A criação de atributos também é chamada de engenharia de atributos.
O BigQuery ML oferece as seguintes técnicas de pré-processamento de atributos:
Pré-processamento automático. O BigQuery ML realiza o pré-processamento automático durante o treinamento. Para mais informações, consulte Pré-processamento automático de atributos.
Pré-processamento manual. É possível usar a cláusula TRANSFORM na instrução CREATE MODEL para definir o pré-processamento personalizado usando funções de pré-processamento manual.
Também é possível usar essas funções fora da cláusula TRANSFORM para processar dados de treinamento antes de criar o modelo.
Receber informações do atributo
Além disso, é possível usar a função ML.FEATURE_INFO
para recuperar as estatísticas de todas as colunas de atributos de entrada.
[[["Fácil de entender","easyToUnderstand","thumb-up"],["Meu problema foi resolvido","solvedMyProblem","thumb-up"],["Outro","otherUp","thumb-up"]],[["Difícil de entender","hardToUnderstand","thumb-down"],["Informações incorretas ou exemplo de código","incorrectInformationOrSampleCode","thumb-down"],["Não contém as informações/amostras de que eu preciso","missingTheInformationSamplesINeed","thumb-down"],["Problema na tradução","translationIssue","thumb-down"],["Outro","otherDown","thumb-down"]],["Última atualização 2025-01-07 UTC."],[[["\u003cp\u003eFeature preprocessing, encompassing both feature creation (engineering) and data cleaning, is a crucial step in the machine learning process.\u003c/p\u003e\n"],["\u003cp\u003eBigQuery ML offers automatic preprocessing during training, simplifying the process for users.\u003c/p\u003e\n"],["\u003cp\u003eManual preprocessing is also available in BigQuery ML, allowing for custom preprocessing definitions using the \u003ccode\u003eTRANSFORM\u003c/code\u003e clause and specific functions.\u003c/p\u003e\n"],["\u003cp\u003eThe \u003ccode\u003eML.FEATURE_INFO\u003c/code\u003e function enables users to retrieve statistics about the input feature columns.\u003c/p\u003e\n"],["\u003cp\u003eBasic knowledge of the ML development lifecycle, including feature engineering and model training, is recommended for better optimization of data and models.\u003c/p\u003e\n"]]],[],null,["# Feature preprocessing overview\n==============================\n\n*Feature preprocessing* is one of the most important steps in the machine\nlearning lifecycle. It consists of creating features and cleaning the training\ndata. Creating features is also referred as *feature engineering*.\n\nBigQuery ML provides the following feature preprocessing techniques:\n\n- **Automatic preprocessing** . BigQuery ML performs automatic\n preprocessing during training. For more information, see [Automatic feature\n preprocessing](/bigquery/docs/reference/standard-sql/bigqueryml-auto-preprocessing).\n\n- **Manual preprocessing** . You can use the [`TRANSFORM` clause](/bigquery/docs/reference/standard-sql/bigqueryml-syntax-create#transform)\n in the `CREATE MODEL` statement to define custom preprocessing using [manual\n preprocessing\n functions](/bigquery/docs/manual-preprocessing#types_of_preprocessing_functions).\n You can also use these functions outside of the `TRANSFORM` clause to\n process training data before creating the model.\n\nGet feature information\n-----------------------\n\nYou can use the [`ML.FEATURE_INFO`\nfunction](/bigquery/docs/reference/standard-sql/bigqueryml-syntax-feature) to\nretrieve the statistics of all input feature columns.\n\nRecommended knowledge\n---------------------\n\nBy using the default settings in the `CREATE MODEL` statements and the\ninference functions, you can create and use BigQuery ML models\neven without much ML knowledge. However, having basic knowledge about the\nML development lifecycle, such as feature engineering and model training,\nhelps you optimize both your data and your model to\ndeliver better results. We recommend using the following resources to develop\nfamiliarity with ML techniques and processes:\n\n- [Machine Learning Crash Course](https://developers.google.com/machine-learning/crash-course)\n- [Intro to Machine Learning](https://www.kaggle.com/learn/intro-to-machine-learning)\n- [Data Cleaning](https://www.kaggle.com/learn/data-cleaning)\n- [Feature Engineering](https://www.kaggle.com/learn/feature-engineering)\n- [Intermediate Machine Learning](https://www.kaggle.com/learn/intermediate-machine-learning)\n\nWhat's next\n-----------\n\nLearn about [feature serving](/bigquery/docs/feature-serving) in\nBigQuery ML."]]
