AchemyAPI是一款在线的API,能够为用户提供机器学习的相关服务(自然语言处理,文本分析),包括文本和视觉数据的一些相关处理。
AlchemyAPI提供的功能:
- REST API端点
- 多种主流语言的SDK
- 返回多种数据格式的源数据
- 鲁棒的SLAs(服务级协议)
- 7天24小时电话支持
REST API:
命名实体抽取、文本分类、关系提取、语言检测、关键词提取、情感分析、作者提取、网页清洗、microformats解析,RSS/ATOM Feed检测
- 情感分析(Sentiment Analysis):document level,entity level,keyword level,and so on
例子 document level
例子 entity level
<entity>
<type>Person</type>
<relevance>0.813554</relevance>
<count>13</count>
<text>Moammar Gadhafi</text>
<sentiment>
<type>negative</type>
<score>-0.120216</score>
</sentiment>
</entity>
例子:keyword level
<keyword>
<text>Moammar Gadhafi</text>
<relevance>0.813554</relevance>
<sentiment>
<type>negative</type>
<score>-0.120216</score>
</sentiment>
</keyword>
- 文本分类
analyze your information, assigning the most likely topic category (baseball, mobile phones, etc.).
- 概念标注(Concept Targging):
例如:("My favorite brands are BMW, Ferrari, and Porsche." = "Automotive Industry").
- 实体提取(Entity Extraction):
- Disambiguation(歧异消除)
- Quotations Extraction(引文提取)
identifying:people,companies,organizations,cities,geographic features and so on
例子:
Company
A corporation, association, partnership, or union, that carries on an industrial enterprise.
Examples:
"... search leader Google, Inc. is releasing ..."
"... joined T-Mobile in early 2006 as chief marketing director ..."Person
A specific individual (a human being).
Examples:
"... elected Barack Obama as president ..."
"... went to Mr. Smith's house ..."
- 关键词提取(Keyword Extraction)
- 关系提取(Relation Extraction):Identifying Subject-Action-Object relation
例子:"A determined Bill Clinton traveled to Chicago"
<subject>
<text>A determined Bill Clinton</text>
<entity>
<type>Person</type>
<text>Bill Clinton</text>
</entity>
</subject>
<action>
<text>traveled to</text>
</action>
<object>
<text>Chicago</text>
<entity>
<type>City</type>
<text>Chicago</text>
</entity>
</object>
- Text Extraction(文本提取):extract page text and title information from any web page.
- 人脸识别(Face Detection)
- 图片地址提取(Image Link Extraction):从文本或web page当中提取出图片地址。
- 图像标注(Image Tagging)
- 作者提取(Authors Extraction):从web或文本中提取作者。
- 语言检测(language Detection)
- RSS/ATOM Feed 检测:从web中检测出feed的链接。
- Microformats解析
- 发表日期提取(Publication Date)