Made some analysis to explore Rs sample representativeness from SRDB-V1 to SRDB-V5
SRDB-V5
First draft manuscript ready and shared with coauthors
R session with Darlin and Jason, explained data operation in R using dplyr using some simple examples
Mean for annual (AFM revision)
Revised the AFM manuscript, sent to coauthors, plan to resubmit it next week
Surrogate
Got Lin Zhang's feedback, she said PCA is not a good way, she suggests using cluster
Hi Jinshi,
I don’t think it’s a good idea to use PCA this way, if you don’t have a very prestigious paper coming from a reputable author in your field to do it this way. Even there is such a paper, I still think it’s kinda statistically counterintuitive.
Do I understand it correct that if some models p-values etc contribute most on the same pc, you could say these models are similar? (This part in the paper I didn’t quite get it either).
I kinda recall we had some discussion about model comparison (or something like clustering, not using p-values) a few years ago. Do you have any recollection about it (I know it’s been a long time)?
Best,
Lin
Got Jianqiu's feedback
-- Jianqiu's comments are generally positive, and most comments are very straightforward to address
Task this week
Global C
Find SRDB-V1 to SRDB-V3, updating representativeness analysis
Github (multiple branches need to merge to master, solution?)
SRDB-V5
Work together with Darlin and Jason and have the figure ready for SRDB-V5 manuscript
Finish the 'Site_ID' issue, and will take some time on data collecting
Done last week
Global C
SRDB-V5
Mean for annual (AFM revision)
Surrogate
I don’t think it’s a good idea to use PCA this way, if you don’t have a very prestigious paper coming from a reputable author in your field to do it this way. Even there is such a paper, I still think it’s kinda statistically counterintuitive.
Do I understand it correct that if some models p-values etc contribute most on the same pc, you could say these models are similar? (This part in the paper I didn’t quite get it either).
I kinda recall we had some discussion about model comparison (or something like clustering, not using p-values) a few years ago. Do you have any recollection about it (I know it’s been a long time)?
Best, Lin
Task this week
Global C
SRDB-V5