Dmitry Kaminskiy is the general partner at Deep Knowledge Group, a UK-based consortium of commercial and non-profit organisations active in the realm of deeptech and frontier technologies
Many investors base their investment decisions on data. However, this information—which could be related to an industry’s or company’s performance, its products, the market in which it operates, micro- and macro-economic trends, and more — may actually be insufficient to make a truly informed decision. Inadequate information often leads to poor investment decisions, and therefore, many investors end up losing money.
Additionally, the available data could be weak for several reasons. Consider a scenario in which there are many competitors without clear reasons why a particular competitor is likely to excel compared to others. Or perhaps the investee under consideration might not be sharing full information as it fears creating existential risks to its operations when disclosing certain data. Furthermore, there might be no obligation for a company to disclose any information.
Another reason for bad investment decisions in relation to data — that might seem counterintuitive but makes sense — is this: There is simply too much information available.
In many circumstances, people rely on gut feeling when faced with a decision. It could be something as mundane as selecting a t-shirt to buy or deciding on what to cook for dinner when it is unnecessary to take into account statistics and data analysis. In these cases, relying on your hunch is acceptable and usually produces the desired outcome.
However, while investors weigh up investment options, it is necessary to minimise intuition, set aside emotion, and consider the available facts instead.
Investors are increasingly looking beyond simple linear data. They seek more advanced and sophisticated ways of solving investment problems. This has given rise to data science as an investment tool, representing a multidisciplinary field responsible for extracting insights from data using scientific methods.
The process of objective assessment plays a key role in the strategy around making successful investment decisions. While opinion may vary on what exactly an ‘objective decision’ is and how best to reach it, one thing is indisputable – the process requires the curation and analysis of a lot of data.
Data storage revolution
Over the past couple of decades, there has been a notable evolution in the storage of data, its aggregation, and analysis. But a really remarkable shift happened with the massive implementation of AI for these purposes.
Big data analytics empowered by AI can be used to generate a more accurate and comprehensive assessment of companies’ business performance, as well as their economic, social, and environmental impact. This empowers investors by giving them confidence in their decisions and visibility over the potential effects of their investments, helping them take non-financial factors into consideration without compromising on returns.
Data storage is chiefly the domain of data engineers and data architects, so let us consider the analysis and aggregation of information. How can tonnes of data be analysed effectively? Where can investors locate this data and how do they use it to make optimal investment choices?
Alternative data represents sets of information about a particular company or industry and competitors. This information is published by external sources and can provide important insights into investment opportunities.
Alternative data often provides an astonishing amount of information on market players. In some cases, this may not be the information sought by an investor; however, it might indicate certain features and present important insights. Sentiment analysis is a great example of this and of how it might be used for investment purposes.
Czech data scientists, Petr Hajek and Josef Novotny, who in January 2022 authored a study named ‘Fuzzy Rule-based Prediction of Gold Prices Using News Effect’, devised a highly interpretable trading strategy in terms of rule complexity. Their gold price trend predictions show improved accuracy compared to the number of benchmarks and their model outperforms other state-of-the-art models using alternative baseline strategies in terms of predicting one-day-ahead movement in gold futures price.
Their philosophy is that diverse AI methods for predicting the gold price outperform traditional statistical methods. Previously, research neglected the transparency of legacy systems, which do not incorporate the important effect of media sentiment on investment decisions. They believe their system, by taking into account the news effect, is better than other models which use machine learning and deep learning algorithms. This has important ramifications for investors.
Social sentiment analysis investing
Predicting commodity stock prices with the use of sentiment analysis is in fact an AI technique. Sentiments expressed by reliable sources on social media and other news channels can play a key role in the investment arena.
This poses some critical questions such as 1) How do I choose the right sources for my sentiment analysis?, 2) Is it only applicable for commodities? What about stocks, private companies, and other equities?, and 3) What is behind the curtain? In other words, which model, metrics, or training techniques are used to gather the data?
Taking the guesswork out of media data analysis by using AI can provide fast, granular insights that can help steer critical investment decisions.
Data science can provide integrated media analysis from multiple platforms, giving investors the power to leverage industry- and use-case-specific AI, and eventually apply the insights to steer investments.
While investors do not have to rely on hundreds of sources and an army of AI-driven techniques, access to highly accurate real-time data generated through data science, AI, and big data harvesting, is key to boosting return on investment. However, it is important to supplement this information with insights from alternative sources too, in order to make well-rounded investment choices.