Reputation: 5644
I'm running cor()
on a data.frame
with all numeric values and I'm getting this as the result:
price exprice...
price 1 NA
exprice NA 1
...
So it's either 1
or NA
for each value in the resulting table. Why are the NA
s showing up instead of valid correlations?
Upvotes: 58
Views: 187898
Reputation: 814
In my case I was using more than two variables, and this worked for me better:
cor(x = as.matrix(tbl), method = "pearson", use = "pairwise.complete.obs")
However:
If use has the value "pairwise.complete.obs" then the correlation or covariance between each pair of variables is computed using all complete pairs of observations on those variables. This can result in covariance or correlation matrices which are not positive semi-definite, as well as NA entries if there are no complete pairs for that pair of variables.
Upvotes: 7
Reputation: 105
very simple and correct answer
Tell the correlation to ignore the NAs with use argument, e.g.:
cor(data$price, data$exprice, use = "complete.obs")
Upvotes: 8
Reputation: 31
The NA can actually be due to 2 reasons. One is that there is a NA in your data. Another one is due to there being one of the values being constant. This results in standard deviation being equal to zero and hence the cor function returns NA.
Upvotes: 3
Reputation: 2183
Tell the correlation to ignore the NAs with use
argument, e.g.:
cor(data$price, data$exprice, use = "complete.obs")
Upvotes: 88
Reputation: 3591
The 1
s are because everything is perfectly correlated with itself, and the NA
s are because there are NA
s in your variables.
You will have to specify how you want R to compute the correlation when there are missing values, because the default is to only compute a coefficient with complete information.
You can change this behavior with the use
argument to cor
, see ?cor
for details.
Upvotes: 43
Reputation: 18638
NAs also appear if there are attributes with zero variance (with all elements equal); see for instance:
cor(cbind(a=runif(10),b=rep(1,10)))
which returns:
a b
a 1 NA
b NA 1
Warning message:
In cor(cbind(a = runif(10), b = rep(1, 10))) :
the standard deviation is zero
Upvotes: 18