AI Powered Credit Insights & Analytics

Summary:
Accern generates credit signals based on over a billion public news and social media content for companies, sectors and regions. A credit signal is identified for a company if an article contains mentions of negative news about credit events such as defaults, bankruptcy, covenant default, debt restructuring, rating downgrades, lawsuits, employment actions, analyst rating, downsizing, fraud events and many more, pertaining to a company.

Metrics:
- History from 2014
- Coverage of all US Companies and Sectors
- Over 150 Credit Events (i.e. Bankruptcy, Legal Actions, Analyst Ratings, etc.)
- Companies are mapped using Bloomberg FIGIs and Tickers
**coverage (e.g. private companies) and credit events can be extended if required**

Methodology:
Accern's credit NLP model has been trained on massive amounts of distressed company news and uses Accern's knowledge graph. The model identifies only highly relevant articles and passages that contain discussions of credit risk issues. For each passage, the model then computes a credit sentiment along with a number of other analytics (https://www.accern.com/adaptive-nlp).

Key Fields:
entity_ticker
event_relevance
event_sentiment
event_text
harvested_at
event
event_group
doc_url

A full data dictionary can be found (https://drive.google.com/file/d/1-S7TeQ48KxuGyNKcZT3yf6x-3F2IngU9/view)

Use Cases:
Financial Services institutions (i.e. Banks, Insurance and Asset Management Firms) can use the dataset to help inform them of potential rating migrations, the risk of default and late payments from their service providers and counterparties by quickly being alerted of abnormal exposures based on adverse content around on a company such as default, bankruptcy, debt restructuring, lawsuits etc.

Identifying early warning signals
Detect a credit migration before it occurs by forecast the potential deterioration in the health of a company based on market data and Accern's credit signals.

Improving their research efficiency
By filtering on the credit signals, an analyst eliminates 99% of the noise and improves their initial searches.

搜集汇总

背景与挑战

背景概述

该数据集利用超过十亿条公开新闻和社交媒体内容，为美国公司、行业和地区生成信用信号，涵盖破产、诉讼等150多种信用事件。它基于Accern的NLP模型和知识图谱，帮助金融机构及时预警信用风险并提升研究效率。