Re: P-Sequence, Date-function extraction of event ages, Difference-function e...

[Rayfo...@aol.com]

Hi Philipp,

 

I don't begin to understand what you are trying to achieve, however, if you wish to access the same parameter in different places, then in order ensure it is only calculated once, then the format needs to use the '=' sign

 

For example:

 

   Difference("event2-event1", "event2","event1");
   R_Date("1", 1000, 10)
   {
    z=60;
   };
   Boundary("core top")
   {
   };
  };
  Sum("recurrence times")
  {
   Date("=event5-event4");
   Date("=event4-event3");
   Date("=event3-event2");
   Date("=event2-event1");
  };
 };

 

I stress that this may not be what you are aiming for, but something like it may be appropriate.  Note that here, 'Sum' is not a probability distribution in the usual sense.

 

Hope this helps

 

best wishes

 

Ray

 

 

In a message dated 14/08/2015 17:14:51 GMT Daylight Time, pekky...@gmail.com writes:

Dear fellow OxCal-users,

I work with event deposits in lakes. I am making an age-depth model in the P_Sequence and feed it with my RC date info using the R_Date-function with depth info. I have a couple of events that I want to date, so I include the Date-function for the depths where the events are. In order to also learn about the recurrence time I include the Difference-function on the events. I give a generic example below.

P_Sequence(0.5)
{
Boundary("core base") {z=300;};
R_Date("6", 6000, 60) {z=260;};
 Date("event5") {z=240;};
R_Date("5", 5000, 50) {z=220;};
 Date("event4") {z=200;};
  Difference("event5-event4", "event5","event4");
R_Date("4", 4000, 40) {z=180;};
 Date("event3") {z=160;};
  Difference("event4-event3", "event4","event3");
R_Date("3", 3000, 30) {z=140;};
 Date("event2") {z=120;};
  Difference("event3-event2", "event3","event2");
R_Date("2", 2000, 20) {z=100;};
 Date("event1") {z=80;};
  Difference("event2-event1", "event2","event1");
R_Date("1", 1000, 10) {z=60;};
Boundary("core top") {z=0};
};

This makes a nice model, results in nice event ages, and models the inter-event time very well. So far so good. Now I want to make a sum of all the inter-event times in order to find an overall recurrence pattern. Hoping for the best I tried by appending this:

Sum("recurrence times")
{
 Difference("event5-event4","event5","event4");
 Difference("event4-event3","event4","event3");
 Difference("event3-event2","event3","event2");
 Difference("event2-event1","event2","event1");
};

But that doesn't work. I hope that it becomes clear what I want to do. I didn't find any clue in the manual or in published papers on how to do it. Can it be done in OxCal? And if yes, how?

Cheers,
Philipp

--
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.

[Pekky K]

Hello Ray,
thank you for answering. I have included "=" signs, and it doesn't do it. I understand the general idea behind the "=" sign, and it is more elegant this way, but it is not a fix. You said you don't begin to understand what I want to do, so I assume I wasn't clear enough.

Imagine you have a record with 30 events and say you have plenty of RC dates for OxCal to model the age of each event with a date-function (Date("event1)). You can look at the inter-event period, if you include a difference-function for each inter-event period( Difference("event2-event1","event1","event2")). Now you have 29 inter-event period PDFs and you think you see a regular pattern in these PDFs. Is there a way to sum all inter-event period PDFs to see the overall pattern?

Hope that clarifies it a bit,
Philipp

[MILLARD A.R.]

I don't think Sum works on anything other than dates. So if you want to manipulate the recurrence times you need to do so explicitly using OxCal's algebra. Also as Ray says you should cross-reference the differences if you use them twice, or you can just calculate them outside the P_Sequence. You could calculate the mean recurrence and the sum of recurrence distributions as below, but the sum of distributions does not have any obvious meaning to me.

Plot()
{

P_Sequence(0.5)
{
Boundary("core base"){z=300;};
R_Date("6", 6000, 60){z=260;};
Date("event5"){z=240;};
R_Date("5", 5000, 50){z=220;};
Date("event4"){z=200;};

R_Date("4", 4000, 40){z=180;};
Date("event3"){z=160;};

R_Date("3", 3000, 30){z=140;};
Date("event2"){z=120;};

R_Date("2", 2000, 20){z=100;};
Date("event1"){z=80;};

R_Date("1", 1000, 10){z=60;};
Boundary("core top"){};
};

// recurrence times
Difference("d54","event5","event4");
Difference("d53","event4","event3");
Difference("d52","event3","event2");
Difference("d51","event2","event1");
mean_d = (d51+d52+d53+d54)/4;
sum_d = d51|d52|d53|d54;
};


Best wishes

Andrew
--
 Dr. Andrew Millard 
e: A.R.M...@durham.ac.uk | t: +44 191 334 1147
 w: http://www.dur.ac.uk/archaeology/staff/?id=160
 Senior Lecturer in Archaeology, Durham University, UK

[Pekky K]

Dear Andrew,

thanks for your reply. The sum you mention has indeed no meaning, at least as far as I am concerned. We are talking about two different operations. I mean an average of the density axis (y-axis) of all recurrence time PDFs. The mathematical operation should be somewhat similar to calculating a pooled RC date from multiple RC dates, see fig. 2 in this link:

http://web.pdx.edu/~amesk/pdfs/Radiocarbon%20Dates.pdf

Except that instead of RC dates, I would like to use the modeled recurrence time PDFs as input. How would you go about this in OxCal's algebra.

Cheers,
Philipp

[MILLARD A.R.]

Dear Philipp,

Then that is the mean_d I calculated. The sum_d I calculated is the same operation as the Sum function.

Best wishes

Andrew
--
 Dr. Andrew Millard 
e: A.R.M...@durham.ac.uk | t: +44 191 334 1147
 w: http://www.dur.ac.uk/archaeology/staff/?id=160
 Senior Lecturer in Archaeology, Durham University, UK


> -----Original Message-----
> From: ox...@googlegroups.com [mailto:ox...@googlegroups.com] On Behalf
> Of Pekky K
> Sent: 17 August 2015 13:31
> To: OxCal
> Subject: Re: P-Sequence, Date-function extraction of event ages,
> Difference-function e...
>

[Rayfo...@aol.com]

Hi Andrew, Philipp,

 

Just curious re Philipp's desire to doing 'something similar to pooled RC dates'

 

OxCal doesn't allow the Combine function directly on the Difference values d54, d53, d52, d51.  However if the values from the first table are input directly into a Combine using these values as in the second table here, we get the pooled mean of the Differences (I think).  However the sigma value is double that in the mean_d calculation.

 

(-1299.5 to -1126.5 (95.4%)   (-1213, 43))   vs    (-1388 to -1033 (95.4%)  (-1210, 89)

 

Comment?

 

 

 

 

 

Best wishes

 

Ray