This is a summary of links featured on Quantocracy on Tuesday, 04/14/2020. To see our most recent links, visit the Quant Mashup. Read on readers!
-
Inverting Differentiated Time-Series in pandas for Deep Learning Prediction Analysis [Quant at Risk]A differentiation of the time-series is a common transformation used when we want to get a stationary time-series given a non-stationary one. The latter usually displays time-dependent relationships like trends, seasonality, quasi-cyclic patterns, and their Fourier power spectrum is characterised by the colour noise. On the other hand, stationary time-series summary statistics are not dependent on
-
Trading and investing performance – year six [Investment Idiocy]Time for the annual review post, as my reviews follow the UK tax year which ended on the 5th April. And what a year it has been; well 10 months or so of fairly normal stuff, followed by several weeks of stomach churning market chaos. Previous updates can be found here, here, here, here and here. This post will follow the format of previous posts, but there will be some extra stuff related to the
-
Trend Following Reality: You Need Trends to Trend-Follow [Alpha Architect]Trend Following, as an investing strategy has delivered strong performance during market chaos (e.g., Global Financial Crisis of 20072009), but the strategy has gone through a significant drawdown (save the last few months where things are perking up!). We have seen dismal returns in the recent decade relative to the historical record (this article refers to the period from 2010-2018). In some