AXA Investment Manager’s quantitative team, Rosenberg Equities, is to integrate advanced modelling techniques that utilise neural networks in its investment approach.
The first example of the integration is the use of a neural network model in the sustainable equity strategy, which aims to improve the ability to identify stocks at risk of extreme price events, mitigate tail risk and enhance risk/return outcomes for clients. AXA IM Rosenberg Equities Europe CEO, Gideon Smith, said: “This is the first step in us using neural network techniques and advanced artificial intelligence in client portfolios, but is a natural progression of the advanced quantitative techniques we’ve adopted over the last thirty years.”
The fund manager said it is also looking at “unstructured” data sets to complement traditional financial data used by investors. As an example, it said it is developing a natural language processing model to analyse company filings and other text sources to create new insights into investor sentiment. AXA IM Rosenberg Equities CEO, Heidi Ridley, said: “Both of these developments are a great example of how we can use artificial intelligence and unstructured data to complement our existing modelling techniques, supporting our belief that innovative use of data and technology leads to better insights and ultimately better outcomes for our clients.”