在 Google Cloud 控制台中,Vertex AI 提供一个滑块,用于调整所有类别或标签或者单个类别或标签的置信度阈值。滑块显示在评估标签页的模型详情页面上。置信度阈值是模型为测试项分配类或标签时必须具有的置信度水平。通过调整阈值,您可以看到模型的精度和召回率变化。较高的阈值通常可以提高精确率,并会降低召回率。
[[["易于理解","easyToUnderstand","thumb-up"],["解决了我的问题","solvedMyProblem","thumb-up"],["其他","otherUp","thumb-up"]],[["很难理解","hardToUnderstand","thumb-down"],["信息或示例代码不正确","incorrectInformationOrSampleCode","thumb-down"],["没有我需要的信息/示例","missingTheInformationSamplesINeed","thumb-down"],["翻译问题","translationIssue","thumb-down"],["其他","otherDown","thumb-down"]],["最后更新时间 (UTC):2025-08-28。"],[],[],null,["# Interpret prediction results from text classification models\n\n| Starting on September 15, 2024, you can only customize classification, entity extraction, and sentiment analysis objectives by moving to Vertex AI Gemini prompts and tuning. Training or updating models for Vertex AI AutoML for Text classification, entity extraction, and sentiment analysis objectives will no longer be available. You can continue using existing Vertex AI AutoML Text models until June 15, 2025. For a comparison of AutoML text and Gemini, see [Gemini for AutoML text users](/vertex-ai/docs/start/automl-gemini-comparison). For more information about how Gemini offers enhanced user experience through improved prompting capabilities, see [Introduction to tuning](/vertex-ai/generative-ai/docs/models/tune-gemini-overview). To get started with tuning, see [Model tuning for Gemini text models](/vertex-ai/generative-ai/docs/models/tune_gemini/tune-gemini-learn)\n\nAfter requesting a prediction, Vertex AI returns results based on your\nmodel's objective. Predictions from multi-label classification models return one\nor more labels for each document and a confidence score for each label. For\nsingle-label classification models, predictions return only one label and\nconfidence score per document.\n\n\nThe confidence score communicates how strongly your model associates each\nclass or label with a test item. The higher the number, the higher the model's\nconfidence that the label should be applied to that item. You decide how high\nthe confidence score must be for you to accept the model's results.\n\n\u003cbr /\u003e\n\nScore threshold slider\n----------------------\n\n\nIn the Google Cloud console, Vertex AI provides a slider that's\nused to adjust the confidence threshold for all classes or labels, or an\nindividual class or label. The slider is available on a model's detail page in\nthe **Evaluate** tab. The confidence threshold is the confidence level that\nthe model must have for it to assign a class or label to a test item. As you\nadjust the threshold, you can see how your model's precision and recall\nchanges. Higher thresholds typically increase precision and lower recall.\n\n\u003cbr /\u003e\n\nExample batch prediction output\n-------------------------------\n\nThe following sample is the predicted result for a multi-label classification\nmodel. The model applied the `GreatService`, `Suggestion`, and `InfoRequest`\nlabels to the submitted document. The confidence values apply to each of the\nlabels in order. In this example, the model predicted `GreatService` as the most\nrelevant label.\n\n\n| **Note**: The following JSON Lines example includes line breaks for\n| readability. In your JSON Lines files, line breaks are included only after each\n| each JSON object.\n\n\u003cbr /\u003e\n\n\n```\n{\n \"instance\": {\"content\": \"gs://bucket/text.txt\", \"mimeType\": \"text/plain\"},\n \"predictions\": [\n {\n \"ids\": [\n \"1234567890123456789\",\n \"2234567890123456789\",\n \"3234567890123456789\"\n ],\n \"displayNames\": [\n \"GreatService\",\n \"Suggestion\",\n \"InfoRequest\"\n ],\n \"confidences\": [\n 0.8986392080783844,\n 0.81984345316886902,\n 0.7722353458404541\n ]\n }\n ]\n}\n```"]]