For many decades, there has been interest in developing computer algorithms that improve automatically through experience or “machine learning” (ML). Arthur Lee Samuel, a pioneer in computer gaming and artificial intelligence, developed the first program able to improve its own performance at playing checkers/draughts game and in 1962 it defeated a human player. Machine Learning was born.
Since then, research on ML algorithms, coupled with increased available computer power and available data has allowed addressing more complex problems, breaking milestones of performance on tasks that were once deemed either too complex or simply out of reach for non-human “intelligent” systems. Unlike in the early days of ML, today’s state of the art ML algorithms are able to “learn” by processing huge amounts of data, extracting the relevant information out of it.
At Adaptive, we have the technical expertise to design, build and operate bespoke trading venues implementing different types of market models: central limit order books (CLOB) are a great fit for mature liquid markets with many participants, request for quotation (RFQ), indication of interest (IOI) and variations of these which support the discovery of liquidity in less-liquid cases. Regardless of the market model, however, the trading venues we deliver allow our clients to retain ownership of the data it generates. We can help them understand the behaviour of the venue and its market participants through the analysis of this data, as well as design and train ML predictive models leveraging this data.
The following two articles are about ML models in finance. The first article reviews the history of ML and the different types of ML models, with a focus on their application in finance: banking (analyze data generated by their customers, their financial transactions or the reports filed by companies, target marketing campaigns, assess credit default risk, fraud detection), asset management (portfolio optimization, risk control, detect market regime changes, analyze correlations amongst assets) and trading (price prediction, market impact and liquidity assessment, algorithmic trading). Published ML models in finance more frequently focus on price prediction (~79%) tasks. Price is one of the factors that affect the returns of an investment, and a difficult one to predict, but it’s not the only factor. There are also the costs of trading, some of which are explicit and can be calculated before trading (taxes, fees), but others (like market impact) are implicit and have to be predicted.
The second article focuses on market impact models, how placing and executing orders affect the price. It describes how these models have been modified to provide predictions that better fit the observed market impact, and hint at some ML algorithms which could further improve the quality of the predictions. Predicting market impact cost helps fund managers, investment banks and traders choose a trading strategy without unaccounted and unexpected market impact costs. It suggests ways in which this prediction could be improved: train ML algorithms on larger datasets with more than a few input variables, train neural networks designed for sequential data like financial time series (long short term memory, recurrent and some types of convolutional neural networks) to include previous prices in the input to the algorithm.
Adaptive can help develop such models using the data that flows in the trading applications that we develop for our clients. Reach out to us if you have data analysis needs or would like to leverage trading data to train predictive models.
Abstract "Machine Learning in Finance"
Part I "Machine learning models in finance"
Part II "Market impact of orders, and models that predict it"
Senior Software Developer & Data Scientist,
Adaptive Financial Consulting