Rotation of the loading matrix

[WDZh721]

Hi. I am using function "tenFM.est" to estimate the loading matrices and factor process of a matrix-valued time series. The results of the obtained loading matrix do not seem to have been rotated using the varimax procedure, making it difficult to interpret the importance of the variables.

I've checked your source code in file "tenFM.R". Line 23 mentions importing function "varimax", but it doesn't seem to be used later for rotation. I am also curious about the role of x0 in lines 120-122 and whether it is related to normalization or rotation of the estimated results. It doesn't seem to be used elsewhere.

Could you please tell me if the current results of loading matrix of "tenFM.est" have been rotated, or how can I get the rotated matrix. Thank you very much!

[WDZh721]

Hi. I also estimated the autoregressive model for a matrix-valued time series using the "tenAR.est" function, and then found that the obtained standard error for the first coefficient matrix seemed to be problematic.

I've checked the source code of "covtosd" function in file "help.R". In lines 49 and 50, the problem seems can be fixed when I replace sum((dim^2)[1:(k-1)]) and sum((dim^2)[1:k]) with sum((dim^2)[0:(k-1)]) and sum((dim^2)[0:k]), respectively. Could you please help to check if this is correct? Thanks a lot!