Machine Learning

AgPa #26: Trading on Price Charts

(Re-)Imag(in)ing Price Trends (2022)
Jingwen Jiang, Bryan T. Kelly, Dacheng Xiu
The Journal of Finance, Forthcoming, URL

This week’s AGNOSTIC Paper is about technical analysis. Full disclosure: I never believed in technical analysis in the sense of drawing lines on charts or imagining somewhat arbitrary patterns.

But the approach of this week’s authors is quite different. They borrow methodology from image recognition and train a machine learning model to detect predictive patterns in price charts (Yes, the machine receives the price chart as picture, not the underlying numbers!)…

  • The model identifies very profitable short-term signals
  • The signals are also profitable over longer horizons
  • Some of the machine-learning-signals are explainable
  • The model disagrees with conventional technical analysis

Read the Full Post

AgPa #21: AI-Powered vs. Human Funds

Do AI-Powered Mutual Funds Perform Better? (2022)
Rui Chen, Jinjuan Ren
Finance Research Letters, Volume 47, Part A, URL/SSRN

This week’s AGNOSTIC Paper compares the performance of AI-powered- and human mutual funds between 2017 and 2019 in the US. Although AI-powered funds are not the holy grail some investors may have hoped for, they still added value compared to their human peers…

  • AI-powered mutual funds did not outperform the US market
  • But AI-powered funds outperformed their human peers
  • And AI-powered funds avoided the disposition- and rank effect

Read the Full Post

#3: Big Data & Machine Learning in Asset Management

This week I gave a talk on “Big Data and Machine Learning in Asset Management” at Goethe-University in Frankfurt. Thanks again to my thesis-supervisor Sasan Mansouri for the invitation. In this post I will summarize a few points of the talk and share the slides. The key result is the following framework to evaluate investment strategies that claim to use big data and machine learning. I also apply this to several real world funds.

Read the Full Post