Columns in the output table in Oxcal

206 views
Skip to first unread message

Richard Chiverrell

unread,
Oct 5, 2016, 6:10:08 AM10/5/16
to OxCal

Hi all

Slightly curious as to what the Indices column P displays in model output, essentially during a series of Sequence models run in Outlier Mode I was curious as its function?

On here there is previous discussion of the merits of A (Agreement) and C (Convergence) columns, but what does the P columns does and when/why is it populated (e.g. some times there are numbers sometimes not) I have struggled to find in the help etc.

cheers
Rich

Christopher Ramsey

unread,
Oct 5, 2016, 6:43:35 AM10/5/16
to ox...@googlegroups.com
Richard

The P column appears when samples are in an outlier model - the figure here is just 100% minus the outlier probability - so give the posterior probability that the sample is not an outlier. The other case where it appears is if you flag a sample as an outlier using the Outlier() command with no probability. This is typically done in a sequence model - the value given here then gives the probability that the sample does lie at this point within the sequence (between underlying and overlying layers).

Best wishes

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

Rayfo...@aol.com

unread,
Oct 5, 2016, 9:27:38 AM10/5/16
to ox...@googlegroups.com
Hi Rich,
 
I think the answer goes like this:
 
When you ran in Outlier mode, you would have given an initial outlier probability (say 0.05)  After analysis, you can 'View Outliers' to see the posterior effects as a bar chart.  Say the Outlier probability went from 5% prior probability  to 63% posterior probability, then you would see in Table view column 'P' the value '37' (i.e. 100 minus 63), so the 'P' column is showing the Posterior probability of the model fitting the data.
 
Or it could also be looked on as a measure of the weight of that piece of evidence taken into  consideration in the outlier down weighing analysis.
 
regards
 
Ray
 
In a message dated 05/10/2016 11:10:08 GMT Daylight Time, rchiv...@gmail.com writes:

https://lh3.googleusercontent.com/-b10nm0Z8jBQ/V_TLdJg9WDI/AAAAAAAAAB0/SsgDflcnImg_sgAmmq2zs4iH-AfXhDCuACLcB/s1600/Capture.JPG

Hi all

Slightly curious as to what the Indices column P displays in model output, essentially during a series of Sequence models run in Outlier Mode I was curious as its function?

On here there is previous discussion of the merits of A (Agreement) and C (Convergence) columns, but what does the P columns does and when/why is it populated (e.g. some times there are numbers sometimes not) I have struggled to find in the help etc.

cheers
Rich
Reply all
Reply to author
Forward
0 new messages