The effect that information has on security prices has been at the heart of financial economics research for decades. The general consensus is that superior returns can be achieved with superior information.
“Timely processing of structured and unstructured data and understanding the complex interrelationships embedded in such information is critical for the success of any active investment strategy”
However, monitoring what fuels stock performance is a constant challenge for active investment managers.
Active managers are constantly attempting to find inefficiencies in the markets. They do this by using a unique variety of strategies to build portfolios, based on the understanding and insight of their research staff and the information they are able to process. In today’s world, monitoring the wealth of information presents a unique challenge unlike ever before.
The effectiveness of any active investment strategy critically depends on the ability to process the enormous amount of information that is constantly arriving. A firm is impacted not just by news related to it but by any event or information that can impact its operations or business model; for example, a new product introduction by a competitor or a macroeconomic development that drives demand for the firm’s products. In fact, the number of ways a firm can be impacted is both large and complex, and requires extensive monitoring and interpretation of the information universe. Relevant information may not even contain the firm’s name at all, but could have a second or third order effect on it.
The ability to quickly assess events reported in the news, emerging research, blogs, and more is essential to having an effective investment strategy. However, the incredibly high frequency and heterogeneity of information flow now far outstrips the capacity of investment managers, asset managers, hedge fund managers, quantitative managers and other financial services professionals and their staff to process all the data related to a fund against a complex model systematically and consistently. Achieving real-time intelligence for capital markets cannot be attained without the right technology to help guide the process.
Achieving Real-Time Intelligence
Artificial Intelligence (AI)-powered applications based on advanced computational linguistics and natural language understanding can be used to extract signals based on contextual relevance from thousands of global, regional and local sources including internet URLs, news, research, blogs, social media, internal proprietary documents and content from premium sources. Such AI-enabled solutions are used to interpret millions of global events as they occur and analyze them through comprehensive causal models individualized for each company to develop an outlook. Next, they determine its relevance to one or more firms and rate the impact of the information via a rating scale for severity.
To accurately achieve real-time intelligence, the AI-based analysis should incorporate the following components centered around the ideas of context and relevance:
• Firm-specific operating models that comprise a network of factors that will impact the firm
• A way to classify the information into one or more topics and determine its relevance by examining the firm’s Information Model
• An assessment of the potential impact on the firm by analyzing impact phrases within the information
Natural Language Understanding: The Key Element for Determining Context and Relevance
Understanding context is a multi-faceted challenge. Natural language understanding (NLU) involves applying computational linguistics principles to reverse engineer text back to its fundamental ideas, and realizing how the ideas were connected together to form sentences, paragraphs and the full document. As the natural language text is processed, it needs to be done in the right context which can only be done by focusing on the language structure; not just on the words in the text. The words in many languages can be used in multiple senses, so it is important to disambiguate word senses so their usage in a particular document can be accurately understood. Text documents often use domain specific discourse models, (e.g. legal contracts, news articles, research reports, etc.). There are certain properties of such domain discourse models that should be incorporated in the AI technology in order to enhance NLU.
Many words may also be used as proxies within a document. AI technology must have a way to recognize and understand proxies like “Xerox” for “copy.” In some cases, text in a document may refer to knowledge which is not explicitly part of the text. Humans can understand this with prior knowledge. AI technologies on the other hand have to create a repository of global knowledge that can be retrieved to supplement the document text in order to gain full understanding of its meaning.
Active investment strategies depend on identifying such information inefficiencies, however many currently use research methods which simply cannot process the deluge of the near-real-time information now available in a systematic and efficient fashion.
Timely processing of structured and unstructured data and understanding the complex interrelationships embedded in such information is critical for the success of any active investment strategy. Through AI-enabled applications, the challenge of aggregating and interpreting content from around the world can now be more effectively met to help determine the potential relevance and impact of that information on a company’s intrinsic value.