文本分析 API 的语言和区域支持Language and region support for the Text Analytics API

本文介绍了以下每项操作支持哪些语言:情绪分析、关键短语提取、语言检测和命名的实体识别。This article explains which languages are supported for each operation: sentiment analysis, key phrase extraction, language detection and named entity recognition.

语言检测Language Detection

文本分析 API 可以检测多种语言、变体、方言和一些区域/文化语言。The Text Analytics API can detect a wide range of languages, variants, dialects, and some regional/cultural languages. 语言检测返回一种语言的“脚本”。Language Detection returns the "script" of a language. 例如,对于短语“I have a dog”,它将返回 en 而非 en-USFor instance, for the phrase "I have a dog" it will return en instead of en-US. 唯一的特例是中文,对于中文,如果语言检测功能可以确定所提供的文本的脚本,则它将返回 zh_CHSzh_CHTThe only special case is Chinese, where the language detection capability will return zh_CHS or zh_CHT if it can determine the script given the text provided. 当无法识别中文文档的具体脚本时,它将简单地返回 zhIn situations where a specific script cannot be identified for a Chinese document, it will return simply zh.

我们不会发布此功能的确切语言列表,但它可以检测各种语言、变体、方言和一些区域/文化语言。We don't publish the exact list of languages for this feature, but it can detect a wide range of languages, variants, dialects, and some regional/cultural languages.

如果内容是用较少使用的语言表示的,则可以尝试“语言检测”来查看它是否返回代码。If you have content expressed in a less frequently used language, you can try Language Detection to see if it returns a code. 无法检测的语言的响应为 unknownThe response for languages that cannot be detected is unknown.

情绪分析、关键短语提取和命名的实体识别Sentiment Analysis, Key Phrase Extraction, and Named Entity Recognition

对于情绪分析、关键短语提取和实体识别,所支持语言的列表更具选择性,因为分析器已优化为适应更多语言的语言规则。For sentiment analysis, key phrase extraction, and entity recognition, the list of supported languages is more selective as the analyzers are refined to accommodate the linguistic rules of additional languages. 在命名实体识别 v2 中,对全套实体类型的支持目前仅限于以下语言:In Named Entity Recognition v2, support for the full set of entity types is currently limited to the following languages:

  • 英语English
  • 简体中文Chinese-Simplified
  • 法语French
  • 德语German
  • 西班牙语Spanish

对于其他语言,仅返回 PersonLocationOrganization 命名实体。Only the Person, Location and Organization named entities are returned for the other languages.

语言列表和状态Language list and status

语言支持最初在预览版中推广,渐变到正式版 (GA) 状态,且各种语言彼此独立并且总体上独立于文本分析服务。Language support is initially rolled out in preview, graduating to generally available (GA) status, independently of each other and of the Text Analytics service overall. 即使文本分析 API 转变为正式版时,有些语言也可能保持在预览版。It's possible for languages to remain in preview, even while Text Analytics API transitions to generally available.

语言Language 语言代码Language code 情绪Sentiment 关键短语Key phrases 命名实体识别Named Entity Recognition 实体链接Entity linking 注释Notes
阿拉伯语Arabic ar ✔ *✔ *
捷克语Czech cs ✔ *✔ *
简体中文Chinese-Simplified zh-hans ✔ **✔ ** zh 也接受zh also accepted
繁体中文Chinese-Traditional zh-hant ✔ **✔ **
丹麦语Danish da ✔ *✔ * ✔ *✔ *
荷兰语Dutch nl ✔ **✔ ** ✔ *✔ *
英语English en ✔ **✔ ** ✔ **✔ ** ✔ **✔ **
芬兰语Finnish fi ✔ *✔ * ✔ *✔ *
法语French fr ✔ **✔ **
德语German de ✔ **✔ **
希腊语Greek el ✔ *✔ *
匈牙利语Hungarian hu ✔ *✔ *
意大利语Italian it ✔ **✔ ** ✔ *✔ *
日语Japanese ja ✔ **✔ ** ✔ *✔ *
朝鲜语Korean ko ✔ **✔ ** ✔ *✔ *
挪威语(博克马尔语)Norwegian (Bokmål) no ✔ *✔ * ✔ *✔ * nb 也接受nb also accepted
波兰语Polish pl ✔ *✔ * ✔ *✔ *
葡萄牙语(葡萄牙)Portuguese (Portugal) pt-PT ✔**✔** ✔ *✔ * pt 也接受pt also accepted
葡萄牙语(巴西)Portuguese (Brazil) pt-BR ✔ *✔ *
俄语Russian ru ✔ *✔ * ✔ *✔ *
西班牙语Spanish es ✔**✔** ✔ *✔ * ✔ **✔ **
瑞典语Swedish sv ✔ *✔ * ✔ *✔ *
土耳其语Turkish tr ✔ *✔ * ✔ *✔ *

* 语言支持为预览版* Language support is in preview

** 命名实体识别实体链接均适用于此语言。** Named Entity Recognition and Entity linking are both available for this language.

另请参阅See also

认知服务文档页面 Cognitive Services Documentation page
认知服务产品页面Cognitive Services Product page