Launching new Natural Language Processing algorithms to automate sentiment analysis in earnings calls
Introducing four proprietary Machine Learning models to decode valuation drivers of listed companies
Iridium Quant Lens data, analytics and AI to level playing field between companies, analysts and investors
Iridium Advisors, the investor relations consultancy, today unveils two self-funded technologies that decode valuation drivers of listed companies and automate sentiment analysis in earnings calls. Based on algorithms that analyse massive volumes of publicly available data, these innovative tools are part of Iridium’s venture into artificial intelligence, data science and advanced analytics known as “Iridium Quant Lens”.
Oliver Schutzmann, CEO of Iridium Advisors, said: “Iridium Quant Lens is one way we are evolving to support the corporate world to use artificial intelligence to address possible vulnerabilities and unlock potential for improved valuations. By providing organizations and leaders with this cutting-edge technology and pairing it with our knowledge and expertise of investor relations, we equip our clients with an unrivalled level of insights that they can use to close the gap between their business value and their market value.”
Capital markets are a complex, dynamic and nuanced ecosystem that essentially seeks to synthesize a huge array of fundamental and alternative data to make intelligent investment decisions. Institutional investors and research analysts have been quick to embrace Machine Learning to augment their investment process and Natural Language Processing to identify hidden sentiment in earnings calls to confirm or discredit their investment views. Due to the limited availability of affordable tools, public companies across all sectors and geographies have been much slower to integrate these artificial intelligence concepts into their corporate decision-making and investor communications processes.
In a first foray to level the playing field, Iridium Quant Lens now provides Natural Language Processing and Machine Learning algorithms that not only help boards and management teams better understand how these technologies are applied by the investment community, but also to enable public companies to run their own algorithms that can quantify valuation drivers and anticipate market reactions to strategic decisions, financial results and management commentaries after earnings calls.
Iridium Quant Lens Machine Learning algorithms identify to a high degree of accuracy the financial and non-financial drivers of valuations at a particular point in time. The algorithms, using around 9 million data points compiled from 673 banks globally, proved highly successful in decomposing valuation drivers and, explaining up to 95% of valuation variability for certain time periods. Consequently, Iridium Quant Lens can be utilised to classify overvalued and undervalued banks in a particular geography, and can help quantify addressable downside risk and upside potential. This, by extension, enables senior management to maximize shareholder value and to align their communication and business strategy to the most important valuation drivers recognized by Iridium Quant Lens. The algorithms have also highlighted a series of consistently significant valuation patterns which confirm the impact of non-financial metrics such as index inclusion, investor relations activity and credit ratings on share price.
Iridium Quant Lens Natural Language Processing algorithms automate earnings call analysis by quantifying language at a scale and speed that is impossible to replicate by the human brain. To date, it has processed 3 million words from more than 500 earnings call transcripts, representing over 75% of the market capitalization of listed companies across eight stock exchanges in the GCC region. The algorithm generates unbiased insights from the sentiment expressed by management, analysts and investors, the language complexity used, as well as the number of financial metrics conferred during earnings calls. The findings are cross-correlated against analysts’ earnings forecasts to detect management biases, and to predict how changes in sentiment impact short-term market movements.
