Natural Language Processing Enabled Alternative Data
(NLP-AD)

The NLP-AD Team aims to utilize existing and develop new Natural Language Processing (NLP) techniques to extract issue-centric sentiments expressed through the business press and social media as well as in official documents, such as government reports. In addition to English media, the Team will cover Chinese and other media.

The guiding principle is predicated on belief and experiences suggesting that sentiments expressed in texts need to be harnessed with a clear application focus. This issue-centric NLP undertaking has been motivated by the Credit Research Initiative’s (CRI) effort and has developed teaMS to incorporate incremental credit-focused information conveyed in business media on exchange-listed corporations globally.

In addition to complementing the CRI’s global corporate default prediction platform with NLP enabled alternative data, the Team’s efforts will also be directed at:

  • Extracting Greenness Sentiment to support the Green Finance research drive at AIDF and
  • Creating alternative data for SME credit analysis catered to lending institutions.

Professor Duan spoke on the applications of NLP in credit analysis.

    1. Credit-focused NLP_Part I: credit analytics background
    2. Credit-focused NLP_Part II: evidence of incremental performance

In recent years, natural language processing (NLP) has become one of the the hottest field in data science. It is widely used in search, translation, social media monitoring, chat, advertising, recruiting, grammar checking and more. Machine learning makes the results smarter, faster and more precise. Read more:

Back to the Deep Credit Analytics Lab Page