Boost your returns – Join our online course

Boost your returns – Join our online course

The field of machine learning has advanced in leaps and bounds in recent decades. A milestone was passed in 1996 when IBM’s Deep Blue became the first machine to win a chess game against a reigning world champion and, since then, advances in computing power and pattern recognition technologies have seen machines improve their performance at tasks that previously could only be accomplished by humans.

Handwriting, voice and image recognition systems are now well-established, and there are countless other applications finding their way into your life, such as those that can identify fraudulent use of your credit cards; compile lists of books and music you’ll likely enjoy based on your current consumption – even ones that can diagnose medical conditions.

With the advantages of speed and brute force, machines thrive in a world awash with data and complexity, and there is no doubt that the data tide is rising, with important consequences for many industries. Interestingly, last year at Stanford University, Professor Andrew Ng’s CS 229 (Machine Learning) was the most popular course offered. The lecture series is available on YouTube, and over half a million people have watched Lecture 1 (mind you, the course material is a little impenetrable, and the number of views drops sharply for the subsequent lectures).

When it comes to investing, it’s hard to imagine that machines could usurp a skilled analyst who is experienced at delving into a company’s business model, making sense of structural changes that are rippling through an industry, and looking into the eyes management. However, there are some aspects of investing where machines clearly have an edge. For example, if a company’s operating cashflows fall short of its reported NPAT, an analyst might immediately know that’s a cause for concern, but a machine can quickly reference thousands of previous similar cases, and arrive at a good estimate of how big a concern it might be.

At Montgomery, we believe that the best results can be obtained by thoughtfully combining machines and analysts to allow each to focus on that they are best at. The most obvious example of this is our scoring system, which automatically rules out companies that fail our quality and performance requirements, and allows us to focus analyst attention on businesses with sound balance sheets and good economics.

Continual improvement of our processes is part and parcel of long-term success, and recognising recent developments in machine learning, this will be the first of a series of articles where we apply some of the latest pattern recognition technology to sift through the A1-B3 universe and identify a smaller number of investment ideas that really get the machines’ tick of approval. We’ll also highlight the companies that the machines most dislike.

This should give us an interesting set of ideas to explore in more detail, and we’re looking forward to sharing the findings with you as we go. At the outset, it’s important to emphasise the “explore in more detail” point. In some cases, the machines will almost certainly get it wrong, and an important part of the process will be to apply some thoughtful analysis to the short lists to try to separate hidden gems from red herrings.

For the machine learning work, we’ll use something called a Support Vector Machine (SVM). By way of background, SVMs are a fairly recent innovation and arguably the most successful class of machine learning algorithm yet developed. In recent decades they have proven themselves very effective at making sense of large complex data sets, and have come to be used across a wide range of real-world pattern recognition problems.

To use the SVM, we first need to train it. We do this by gathering tens of thousands of historical training examples, each of which contains the financial details of an ASX-listed company at a point in time, together with the investment performance subsequently delivered by that company. In the training process, the SVM analyses these examples and learns the relationships between individual pieces of financial data, and investment performance.

Having being trained in this way, the SVM can then take the financial data for a company it hasn’t seen before, and provide a prediction for the future investment performance of that company.

We are currently preparing the large volumes of historical information needed for the training part of the process. In future updates, we will review some of the insights that emerge from the training exercise, and look at the specific companies that seem to offer the most intriguing investment prospects.

Watch this space for further updates, and in the meantime, feel free to comment below and share your thoughts and questions.

MORE BY TimINVEST WITH MONTGOMERY

Tim joined Montgomery in July 2012 and is a senior member of the investment team. Prior to this, Tim was an Executive Director in the corporate advisory division of Gresham Partners, where he worked for 17 years. Tim focuses on quant investing and market-neutral strategies.

This post was contributed by a representative of Montgomery Investment Management Pty Limited (AFSL No. 354564). The principal purpose of this post is to provide factual information and not provide financial product advice. Additionally, the information provided is not intended to provide any recommendation or opinion about any financial product. Any commentary and statements of opinion however may contain general advice only that is prepared without taking into account your personal objectives, financial circumstances or needs. Because of this, before acting on any of the information provided, you should always consider its appropriateness in light of your personal objectives, financial circumstances and needs and should consider seeking independent advice from a financial advisor if necessary before making any decisions. This post specifically excludes personal advice.