I have two sets of independent variables. I would like to find out the strength and direction of association between these two variables. Since both the variables are independent, and I need to account for error on both the axis, should I use Deming regression on them? Also,I expect that bottom 10 - 20 % of the data is noise which can reduce the strength of the association. Please suggest me the code to automatically remove 10 or 20 % of the data from statistical analysis.
Asked
Active
Viewed 103 times
0
-
You have two separate questions, a statistics question and a programming question. The statistics question belongs on stats.stackexchange, not on Stack Overflow. The programming question is on-topic here, but it is low-quality. You should do some searching like "how to subset data" or "random data subset" and you should find [plenty](https://stackoverflow.com/q/12032307/903061) of [results](https://stackoverflow.com/q/19861775/903061) - if those answers don't work for you, please provide enough of an example to show what you tried and why it was different than you hoped. – Gregor Thomas Jul 27 '17 at 13:34
-
If you need help making good examples, see [How to make a great R reproducible example](https://stackoverflow.com/q/5963269/903061) and [How to Create a Minimal, Complete, Verifiable Example](https://stackoverflow.com/help/mcve). – Gregor Thomas Jul 27 '17 at 13:35