Data Science Projects for DSA Candidates - Data Science Central2020-02-20T00:01:18Zhttps://www.datasciencecentral.com/forum/topics/data-science-projects-for-dsa-candidates?groupUrl=dsa-projects&feed=yes&xn_auth=noFor the spurious correlationsâ€¦tag:www.datasciencecentral.com,2014-05-10:6448529:Comment:1681082014-05-10T11:56:00.027ZVincent Granvillehttps://www.datasciencecentral.com/profile/VincentGranville
<p>For the spurious correlations project, you could actually create one variable X with arbitrary but fixed values, and check how it correlates with thousands of simulated variables. The reason being that any set of observations for X has the same probability to occur, under the uniform distribution assumption. This considerably reduces the number of computations, turning a O(n^2) problem with large n, into O(n), from a computational complexity point of view.</p>
<p>For the spurious correlations project, you could actually create one variable X with arbitrary but fixed values, and check how it correlates with thousands of simulated variables. The reason being that any set of observations for X has the same probability to occur, under the uniform distribution assumption. This considerably reduces the number of computations, turning a O(n^2) problem with large n, into O(n), from a computational complexity point of view.</p>