Michael Salk
Managing Director for Media Solutions and Data Distribution, Moody’s Analytics
Decision makers need the best tools to help them make the most accurate, timely decisions in a world full of news. This requires historical knowledge and fresh perspectives using AI and machine learning.
Artificial intelligence (AI) and machine learning technologies have developed greatly over the past two decades allowing data processing and enrichment to happen in an instant, at scale. As these technologies continue to evolve, capabilities are advancing in innovative ways to benefit decision makers. In news processing, these technologies are now facilitating the identification of new trends and topics as they appear in news reports. A recent example of an emerging trend came earlier this year, when NFT, non-fungible tokens, became a widely reported news topic.
Natural language processing engines, such as Moody’s Analytics NewsEdge NLP, are capable of identifying new terms early on as they are gaining momentum in news coverage, as was the case with NFT, then quickly adding these terms as a new topic of interest to assist users tracking these trends.
Through 20 years of news processing, the NewsEdge team has developed a stable, easily explainable, rich source of metadata which customers rely on to feed their own applications and models with minimal latency. AI and NLP capabilities are opening up possibilities to expand learnings and apply new models to news that inform decision making.
A hybrid qualitative-quantitative approach
Moody’s Analytics has a long history of using data inputs to create predictive scores. This has been done, for example, using data sets such as bond prices to predict credit deterioration or improvement.
By applying the same discipline to news, NewsEdge NLP can identify event-based signals. It combines news feeds with the Moody’s Analytics Credit Sentiment Score (CSS) to provide early indications of credit deterioration. By compiling such adverse credit signals from the news, this helps firms assess credit in the loan origination and portfolio risk monitoring process and track unfavourable media.
News is often the primary reason why the market moves, but it can also be the source of information on why the market is moving.
Diverse inputs drive better decisions
The wide analysis of news reporting on a topic is broader and deeper than would be possible from the vantage point of a single news story. A press release may be the first indication an event has happened, but a more comprehensive understanding is provided from analysis of reporting on newswires, local newspapers and industry analysts. The breadth and depth of perspectives provide a more holistic view.
It is possible, using AI and machine learning to process hundreds of thousands of news articles a day, to tag thousands of subjects, industries, sub-industries and events within these stories. A story may be tagged as relating to ‘drug and device regulatory impact’, but it can also be tagged as relevant to ‘clinical trials’ or ‘recalls’.
Going a step further, a story tagged under ‘clinical trials’ could be sub-tagged ‘approval’, ‘initiation of a clinical trial’, ‘each phase of trials’, ‘failed’, ‘halted’ and a number of other trial nodes. This precision adds value to news by enriching the content with actionable signals to help users react to market-impacting events.
Allowing real time decision making
Stories are categorised and tagged for topic, sentiment and event-based signals to allow real-time decision making in sub-seconds. The NLP engines are learning and getting better all the time and news content can be directly plugged into a customer’s decision models, even algorithmic trading models, which are dependent on real-time insights.
The NewsEdge NLP engine provides people with the option to tag their own content sources in a manner consistent with the content provided by the more than 19,000 news sources delivered via their feeds. News is often the primary reason why the market moves, but it can also be the source of information on why the market is moving.
In the future, market participants will look to mine news to inform their actions, and Moody’s Analytics will be there to help them achieve that. The firm plans to further invest in content tagging, natural language processing, machine learning tools, and its own team of news experts.
