PT  - JOURNAL ARTICLE
AU  - Jeffrey M. Bacidore
TI  - COMMENTARY: Reflections on “Cluster Analysis for Evaluating Trading Strategies”
AID  - 10.3905/jot.2018.13.4.130
DP  - 2018 Oct 31
TA  - The Journal of Trading
PG  - 130--131
VI  - 13
IP  - 4
4099  - https://pm-research.com/content/13/4/130.short
4100  - https://pm-research.com/content/13/4/130.full
AB  - Our paper on Cluster Analysis was inspired by our need to group client data by trading strategy, when the data we were provided did not contain any information on trading strategy whatsoever. We ended up relying on a well-known statistical technique, k-means, which surprisingly had not been used widely in trading applications. At the time, non-quant traders were still reluctant to use quantitative techniques, especially black box applications like k-means. Fortunately, a lot has changed since that time, as quants are now using much more sophisticated techniques, like deep learning. And even more important, non-quant traders and business leaders have become much more accepting of such techniques, making it easier for such advanced techniques to be incorporated into trading applications.TOPICS: Statistical methods, legal/regulatory/public policy