Convert to Reandom Variables- Time Series Analysis

#1
Hi,
I have a set of water quality data that are not random but I want use them in a Copula model so I should have random variables (I know it can done with time series analysis). Can anybody help me in finding a journal paper or book that I can use as a guide.

Thank you
 

Dason

Ambassador to the humans
#2
What do you think you mean by random? I don't think you and me are using the same definitions.
 
#3
Copula is to describe the dependence between random variables but my data is a set of daily data which has TREND s.o. told me that I cant use the data with TREND (because it is not random) and I should do some work on it before using in copula
 

Dason

Ambassador to the humans
#4
Data with a trend can most certainly be random. There are procedures that will require you to detrend your data but I would advise you to stop saying that your data isn't random just because there is a trend in it.
 

BGM

TS Contributor
#6
I think OP has not get the definitions for those Jargon yet.

1. You may do some sort of regression analysis for your trend.

2. You can use some time series model to account for the autocorrelations in the series. Traditionally you are referring to a stationary time series so you may want to detrend it first.

3. For copula you want to model the dependency structure for more than one dimension. Not sure your time series is univariate or not.