I. Introduction :
NLP (Natural Language Process, referred to as NLP), is a technology based on artificial intelligence for enterprises and developers of all walks of life to provide for intelligent text analysis and processing of cloud services , is intended to help users efficiently handle text data , and intelligent digital transformation.
Tencent cloud NLP (Natural Language Process, NLP), formerly known as Tencent Wen-chi natural language processing, now after the new upgrade, officially released v1.0 version.
Product depth integration of the internal Tencent (including AI Lab, the information security team, AI platform part, translated Jun and knowledge Wen team self-study, etc.) excellent NLP cutting-edge technology, relying on the massive Chinese corpus cumulative, comprehensive coverage from basic to advanced smart text handling capabilities .
among them,
- Basic Edition includes lexical analysis, syntactic analysis, discourse analysis, vector technology, sentiment analysis, text correction, text classification;
- Premium Edition includes a sensitive word recognition, text audit.
At this stage, to thank partners and new and old customers the confidence and support of the current product free beta, beta proposed deadline is November 1, 2019 .
【Quick Links】
1- Tencent cloud natural language processing Products: https://cloud.tencent.com/product/nlp
2- Tencent cloud Natural Language Processing Product Documentation: https://cloud.tencent.com/document/product/271
3- Tencent cloud natural language processing API documentation: https://cloud.tencent.com/document/product/271/35484
Second, the product features :
Comprehensive coverage of the product from the lexical, syntactic discourse to the level of granularity of NLP and other capabilities .
among them,
- Lexical analysis including smart segmentation, POS tagging, named entity recognition;
- Syntactic analysis includes syntactic dependency analysis, text correction, sentences and other vector;
- Text analysis including sentiment analysis, keyword extraction, text classification, summarization, sensitive word recognition, text audit.
(A), the word-level natural language processing
1- lexical analysis (LexicalAnalysis)
- Provide intelligent segmentation (basic words and phrases), speech tagging, named entity recognition.
- Professional team of data, models, iterative update program to ensure continuous improvement in recognition performance.
- Users simply call the relevant API interface to get to the desired result, without worrying about new words such as discovery, disambiguation, call performance problems lexical analysis.
2- similar words / synonyms (SimilarWords)
- To provide users with synonyms inquiry service.
- Team digging through a massive network-wide data synonyms, and continued on the data, models, etc. iterative update, to ensure that the effect is always synonymous with the times.
- Future users can also provide product-specific data, and we work to create exclusive thesaurus.
Word-level natural language processing features include: word vector (WordEmbedding), word similarity (WordSimilarity) and so on.
(B), sentence-level natural language processing
3- text Error Correction (TextCorrection)
- Enables automatic error correction of the text, that is a sentence or paragraph of typos (do wrong word) for automatic error correction.
- Users only need to provide business data and logs, without paying attention to technical details and update process, you can enjoy their own customized service error correction service, we can not even provide business data, error correction enjoy universal service.
- In the audit office documents, text quality and other intelligent scene wide range of applications.
Sentence-level natural language processing features include: syntactic dependency analysis (DependencyParsing), sentence vector (SentenceEmbedding), sentence similarity (SentenceSimilarity) and so on.
(III), chapter-level natural language processing
4- sentiment analysis (SentimentAnalysis)
- There is a sentiment analysis of product demand service.
- The service can forward the text message to emotion, to evaluate the negative and neutral.
- Has a very important value in business analytics monitoring public opinion, the topic of supervision, reputation analysis.
5- keyword extraction (KeyWordsExtraction)
- Based on keyword extraction platform, extracting a sentence or paragraph in a word embodies critical information, such as news content for users to achieve automatic extraction of keywords, comments, keyword extraction and other basic services.
- The future will also support user-defined dictionaries, improve the extraction effect in the vertical field.
- This feature supports scenarios include news media information extraction, document structure of the financial scene extraction and so on.
6- sensitive word recognition (SensitiveWordsRecognition)
- Recognized text information in the ad, as well as political sensitivity of the information, and returns the corresponding sensitive words.
- Sensitive information can be used for filtering, monitoring public opinion, the UGC review the text data and the like.
Chapter-level natural language processing features include: automatic summary (AutoSummarization), text classification (TextClassification), text review (TextApproval) and so on.
(Four), more natural language processing
Tencent cloud NLP also provides more features, please refer to the API documentation : https://cloud.tencent.com/document/product/271/35484