Finding subset of least correlated time series from a set of time series

External question — Cross Validated Stack Exchange Author: Rocco Original post: https://stats.stackexchange.com/questions/669061 License: CC BY-SA 4.0 — https://creativecommons.org/licenses/by-sa/4.0/ Adaptation: HTML converted to plain text; contact email addresses removed. I have got a set $T = \{t_1, t_2, t_3, ...\}$ that contains some stock prices time series. I am trying to find a way to construct a new set $S \subset T$ that satisfies the following characteristics: Is of predefined size $s < |T|$ The time series that compose $S$ are "as little related as possible between them", meaning that as a group of time series the correlation of each one with the others is the lowest possible. To build such a set i thought of the following algorithm: Calculate the correlation coefficient between all pairs of time series in $T$ , using something like the absolute of Pearson correlation, obtaining a correlation matrix similar to the following: $t_i$ $t_{i+1}$ $...$ $t_i$ 1 $abs(pearson(t_i, t_{i+1}))$ $...$ $t_{i+1}$ 1 $...$ $...$ 1 For every possible subset of $T$ of size $s$ , take the correlation coefficient for every pair of time series in this new subset and average them, obtaining a kind of "correlation index" (let's call it $c$ ) for the subset. If $s = 3$ and the test subset is $\{t_1, t_2, t_3\}$ then the correlation index would be $c = (abs(pearson(t_1, t_2)) + abs(pearson(t_2, t_3)) + abs(pearson(t_3, t_1)))/3 $ Take the subset with the lowest $c$ Now, i am not sure if this can work, because i read that stock prices must be pre-processed before attempting a correlation ( https://stats.stackexchange.com/a/554837/489630 ) and i am generally unsure about the averaging of the Pearson coefficients. Is there a better way to do it? Any other things i should be aware of? Clarification: The time series are the sequence of closing prices of every trading day in a period of 2 years. If it helps, it is also possible to calculate a moving average to "smooth" the curves.