Systematic hedge funds investing in machine learning

Posted by: Anu Chandra on 23-06-2017

Barclays recently released a report on systematic hedge funds that covered the use of machine learning. We thought it might be helpful to share the most salient parts.

1. Rise of systematic hedge funds

Systematic hedge funds are characterised by use of huge datasets from a diverse range of sources, analysis of data to find persistent patterns and encoding trading rules into algorithms to identify patterns. These rules are used to trade securities globally 24 hours a day. Investments include futures and forwards, commodities, market indexes, bonds, currencies and equities.

The strategy most commonly pursued by systematic macro hedge funds is “trend following”. Historic trading data is analysed to identify past trends. Those trends are used to predict future trends and to algorithms that place trades in anticipation of a market rising or falling. It’s also about looking for trends that won’t continue and betting against them. These are known as ‘counter-trend’ strategies and are based on the expectation that a trend will overshoot once it’s passed an average point.

Systematic HF strategies now account for about $500bn in assets, or almost 17% of the total HF industry.

2. Growth in use of big data sources at systematic hedge funds

The amount of data being created annually on a global basis is increasing exponentially. Increases in computing power have also increased our ability to analyse data on a massive scale. This has enabled systematic HF managers to explore new and larger data sets in pursuit for alpha. This has given rise to increased investment in alternative / ‘Big Data’; 54% of systematic managers according to Barclays study are now employing alternative and ‘Big Data’ sources, such as web scraping, social media data, satellite data, credit card data, etc.

3. Machine learning and big data are the two biggest investment areas at systematic hedge funds

62% of systematic managers are combining big data with machine learning.

4. How Ryelore Ai can help systematic hedge funds

Systematic hedge funds are reporting that the big challenge in big data is that it doesn’t come in a clean numerical format that can be plugged straight into a model. A lot of hard work is needed to get value out of it. It can also be expensive to find and scrub data. At Ryelore Ai, we are experts in applying deep learning to complex vision datasets such as satellite and drone imagery. We as a team combine hedge fund, machine learning and technology expertise. This means that we can do the hard work of applying our skills to complex satellite data so it can be easily plugged straight into a model.