Working with Large Covariance Matrixes

Devon McCormick speaks on data analysis using the beautiful and arcane J language.

Abstract: We will look at using covariance matrixes of stock returns to forecast future returns and some ways of dealing with the large search space entailed by this approach.

Background: The J programming language is a rethinking of APL within the strictures of an ASCII character set. It is extremely terse and features powerful handling of arrays and a ubiquitous form of indirection called "boxing". Documentation of the syntax is curious in that it makes appeal to natural-language concepts such as "verbs", "adverbs", "gerunds", "passives", "infixes", and so on.

Although J itself is a fairly rarefied item, covariance matrixes are not, and their handling may be of interest to a variety of people. Devon McCormick is an excellent speaker and the presentation should be captivating.

Venue: Be aware that the venue is small. Building Security sternly requires us to supply our visitors' actual names a full day in advance, and when you arrive you can expect to be asked by them for ID that shows the name you gave us. When they admit you, they will take your picture for a paper badge they hand you before sending you to the elevators.