Description Usage Arguments Value Examples
Initial data screening suggested that most data did not need to be differenced seasonally first. Simulations show that structure in residuals for one NTU only does not affect type I error. However, structure in two NTU’s can greatly affect type I error. Check autocorrelation functions for undifferenced modified data (as before). Look at row and column sums for each NTU. Most NTUs have only a few acf’s with structure. Most of the acf’s with structure are found in the top 5 upper 5 more likely to be OK. This is a two part script with commands in between to run manually. Determine an appropriate cutoff (I used 5 script is marked by comments below.
1 | s_matrixacf2(matrix, boolean, fn)
|
inputParameter1 |
matrix of NTU abundances in rows and sample dates in columns |
inputParameter2 |
boolean is a matrix of 1's and 0's from the previous linear model code indicating which NTUs are connected (1's) |
inputParameter3 |
fn is a file name for the output of the first function |
inputParameter4 |
fn2 is a file name for the output of the second function |
inputParameter5 |
X is the number of connections at the cutoff |
output A boolean matrix of connections that fail (1’s). Results are in the upper triangle. Need to subtract this matrix from the matrix of connections.
1 | Run diagnostic with this command: s_matrixacf2(matrix=, boolean=,fn=)
|
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.