Error in rowSums: 'x' must be numeric

[Hugh Robinson]

Hi,

I'm trying to run a relatively simple single season occupancy model using occu but get an  "Error in rowSums(y, na.rm = TRUE) : 'x' must be numeric".  My unmarkedFrame seems to be entering properly, summary(x) provides the following.

168 sites
Maximum number of observations per site: 76
Mean number of observations per site: 13.54
Sites with at least one detection: 32

Tabulation of y observations:
    0     1  <NA>
 2195    79 10494

I have a very unbalanced sample design, and I've included several unsurveyed sites in which I would like to predict occupancy based on covariates. 
Could the abundance of NA values in my data be causing the row Sum error?

Thanks,
HR

[Jeffrey Royle]

hi Hugh,

 To test that out, delete the all-NA observations and see what happens, I'm guessing yes.

 You shouldn't need to carry around those missing records in the fitting stage of the analysis (but predict() will work on whatever input covariate values you provide).

regards

andy

 

--
You received this message because you are subscribed to the Google Groups "unmarked" group.
To unsubscribe from this group and stop receiving emails from it, send an email to unmarked+u...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.

[Hugh Robinson]

Thanks Andy,
This also worked.  Ran before creating my unmarkedFrameOccu file, and where detect is the name of my detection history data frame.
detect<-apply(detect, 2, function(x){as.numeric(as.character(x))})

HR

[Brian Leo]

Hi Andy,

I get the same error when trying to run an occupancy model.  I am trying to practice, and have downloaded the data from MacKenzie et al 2002.  I am trying to run the occupancy models for the american toad.  There are many "-" or NA values in the dataset.  When I run the summary of my unmarkedFrameOccu where "data" is the dataframe of occupancy data and habitat is a site covariate vector.

umf <- unmarkedFrameOccu(y = data, siteCovs = data.frame(habitat = habitat))

summary(umf)

28 sites

Maximum number of observations per site: 82 

Mean number of observations per site: 82 

Sites with at least one detection: 10 

Tabulation of y observations:

   -    0    1 <NA> 

2044  221   31    0 

Site-level covariates:

       X1        

 Min.   :0.0000  

 1st Qu.:0.0000  

 Median :0.0000  

 Mean   :0.2857  

 3rd Qu.:1.0000  

 Max.   :1.0000 

When I delete the NAs, as was suggested for Hugh's problem, I get the error when I try to create the unmarkedFrameOccu:

Error in validObject(.Object) : 

  invalid class “unmarkedFrame” object: siteCovData does not have same size number of sites as y.

What is the proper way of handling NAs?

Thanks,

Brian

[Ian Jenkins]

Thank you!! This worked for me. RStudio 'research methods II' 

April 2020

To unsubscribe from this group and stop receiving emails from it, send an email to unma...@googlegroups.com.