About calculating the p value of PD and PE

[Wang]

Hi, Shawn
I want to calculate the p-value of PD. I found the Randomisation button. I would like to ask you how to set it up further. I have reached this step now.

[Wang]

[Shawn Laffan]

Hello Wang,

For most uses you can set the number of iterations to 999 and press the go button.  That will generate 999 random realisations. The significance scores are derived from the rank relative positions of the original values relative to those generated using the random realisations. 

999 iterations are used to allow for better precision around the thresholds.  e.g. for alpha=0.05 one cannot know if the value is 0.049 or 0.051 with only 99 iterations. 

There are some details in section 5.2 of the Quick Start Guide but there could be more.

There are also several relevant blog posts that will help, and a section of the (very incomplete) help system.
https://biodiverse-analysis-software.blogspot.com/search/label/randomisations
https://github.com/shawnlaffan/biodiverse/wiki/AnalysisTypes#where-do-the-randomisation-results-go-and-what-do-they-mean


In terms of the p-scores, the rank relative position of each index is stored under the p_rank lists.

For example, for a randomisation analysis called "rand1" with indices in the SPATIAL_RESULTS list you would see a list called rand1>>p_rank>>SPATIAL_RESULTS next time you open the spatial or cluster analysis. If you run calculations that generate one or more list results then you would see those also, e.g. if you calculate the phylogenetic diversity terminal node list calculation then you would see rand1>>p_rank>>PD_INCLUDED_TERMINAL_NODE_LIST
(see https://github.com/shawnlaffan/biodiverse/wiki/Indices#phylogenetic-diversity-terminal-node-list )

These lists have the same index names as the original lists but the values are the rank relative positions as percentiles.

If you are looking for significantly high values on a one-tailed test then you would look for p_rank values greater than 0.95 for each index.  For a two tailed test with alpha=0.05 you would look for values <0.025 or >0.975.

The actual thresholding needs to be done outside Biodiverse as then you can use whatever alpha cutoff is relevant to your data.  Since Version 4.3, though, Biodiverse colours the values based on significance threshold.  This makes it easier to see which locations are significantly different from the random generation process. 
https://biodiverse-analysis-software.blogspot.com/2023/04/changes-to-randomisation-results-p-rank.html


Regards,
Shawn.

--
You received this message because you are subscribed to the Google Groups "Biodiverse Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to biodiverse-use...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/biodiverse-users/2f4dda6d-393d-4060-8ed1-9cf7758a69f4n%40googlegroups.com.

[Wang]

Hello Shawn,
Thank you very much. My question has been solved. I have another question to ask you. When I was reading the article Phylogenetic measures of biodiversity and neo- and paleo-endemism in Australian Acacia, I didn’t understand the Categorical analysis of neo. - and paleo- endemism, what is the threshold corresponding to this legend and how is it measured? When calculating CANAPE, is the final result a measure of neo- and paleo-endemism based on p-values?

[Wang]

There is another question, how to set up the Super class?

在2023年11月26日星期日 UTC+8 05:50:37<Shawn Laffan> 写道：

[Shawn Laffan]

Hello Wang,

The CANAPE process is also described under these links:
https://biodiverse-analysis-software.blogspot.com/2014/11/canape-categorical-analysis-of-palaeo.html
https://biodiverse-analysis-software.blogspot.com/2014/11/do-it-yourself-canape.html
https://biodiverse-analysis-software.blogspot.com/2022/10/biodiverse-now-calculates-canape-for-you.html

The full set of CANAPE related posts is at https://biodiverse-analysis-software.blogspot.com/search/label/CANAPE

Regards,
Shawn.

To view this discussion on the web visit https://groups.google.com/d/msgid/biodiverse-users/8a0bec0f-2349-435f-9d6b-393945c8bb35n%40googlegroups.com.

[Wang]

Hello Shawn,
I'm sorry that I couldn't find an effective way to add the Super class. I may be too stupid. Professor, can you tell me in detail how to add the Super class? Thank you very much.