Syntax for Bootstrapping
Locked
 
|
Good Friday afternoon, all.
I know that SPSS provides bootstrapping solutions in the Bootstrapping Option or the Premium Edition. However, only the BCa and percentile methods are available. I'm in the process of developing syntax to bootstrap a number of statistics. The link to the first of these is below. This syntax provides solutions for the jackknife and bootstrap analysis of a variable mean. It contains syntax for seven approaches to bootstrap computation of confidence limits - normal, bootstrap-t, ordinary percentile, asymptotic adjusted, hybrid, bias corrected, and bias corrected and accelerated. There is also a solution for the Bayesian bootstrap, which uses the Haldane prior and Dirichlet posterior distributions of the values of the dataset rather than bootstrapping the values themselves. Graphs of the jackknife, bootstrap, and Bayesian bootstrap distributions are produced together with kernel density curves. I am working on an ABC solution, which is an empirical approximation of the BCa.
Next up...the median.
Have a nice weekend.
https://drive.google.com/drive/folders/14tOFH4ViFpGZ78gZefoOFWKLt6hcrDC8?usp=sharing
Brian Dates
|
|
|
Hi Brian,
Thanks for your work ! By any chance, would you be able to develop a syntax to generate bootstrapped confidence intervals for correlations ? The IBM Bootstrapping "add-on" is 696$/year, which is incredibly expensive, so it would be fantastic if someone could come up with a syntax... I'm curious to know whether this would be feasible. Best, MO. -- Sent from: http://spssx-discussion.1045642.n5.nabble.com/ ===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD |
|
|
It's in my plans. I'm adding some detail to the bootstrapping based on DiCiccio and Efron, which presents Length (length between the endpoints of the confidence interval) and Right-Left Asymmetry (the ratio of
the distance from the upper limit of the confidence interval to the mean value minus the distance from the lower limit of the confidence interval to the mean). It's an indicator of the degree of asymmetry of the bootstrap confidence interval. I should have
that done by tomorrow and post the revision. Then I'll get started on the correlations. My plan was to do the median and then the correlation, but it's easy for me to switch that. Should be by the end of the week.
Brian
From: SPSSX(r) Discussion <[hidden email]> on behalf of Oliver <[hidden email]>
Sent: Monday, June 4, 2018 1:18:45 AM To: [hidden email] Subject: Re: Syntax for Bootstrapping Hi Brian,
Thanks for your work ! By any chance, would you be able to develop a syntax to generate bootstrapped confidence intervals for correlations ? The IBM Bootstrapping "add-on" is 696$/year, which is incredibly expensive, so it would be fantastic if someone could come up with a syntax... I'm curious to know whether this would be feasible. Best, MO. -- Sent from: http://spssx-discussion.1045642.n5.nabble.com/ ===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD |
Administrator
|
http://spssx-discussion.1045642.n5.nabble.com/Sampling-WITH-replacement-td5618318.html#a5620719
Perhaps useful? ----- Please reply to the list and not to my personal email. Those desiring my consulting or training services please feel free to email me. --- "Nolite dare sanctum canibus neque mittatis margaritas vestras ante porcos ne forte conculcent eas pedibus suis." Cum es damnatorum possederunt porcos iens ut salire off sanguinum cliff in abyssum?" -- Sent from: http://spssx-discussion.1045642.n5.nabble.com/ ===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD
Please reply to the list and not to my personal email.
Those desiring my consulting or training services please feel free to email me. --- "Nolite dare sanctum canibus neque mittatis margaritas vestras ante porcos ne forte conculcent eas pedibus suis." Cum es damnatorum possederunt porcos iens ut salire off sanguinum cliff in abyssum?" |
|
|
David,
Thanks!
Brian
From: SPSSX(r) Discussion <[hidden email]> on behalf of David Marso <[hidden email]>
Sent: Monday, June 4, 2018 1:17:40 PM To: [hidden email] Subject: Re: Syntax for Bootstrapping http://spssx-discussion.1045642.n5.nabble.com/Sampling-WITH-replacement-td5618318.html#a5620719
Perhaps useful? ----- Please reply to the list and not to my personal email. Those desiring my consulting or training services please feel free to email me. --- "Nolite dare sanctum canibus neque mittatis margaritas vestras ante porcos ne forte conculcent eas pedibus suis." Cum es damnatorum possederunt porcos iens ut salire off sanguinum cliff in abyssum?" -- Sent from: http://spssx-discussion.1045642.n5.nabble.com/ ===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD |
|
|
In reply to this post by Oliver
Not to denigrate Brian's work at all, but I should point out that the Bootstrapping option is included in the Standard system in SPSS Subscription. In the regular V25, the Premium version includes everything in Professional plus Bootstrapping, Direct Marketing, Complex Samples, Neural Networks, Conjoint, and Exact Tests. See for details. IMO, Bootstrapping ought to be in the Standard version, as it is for SPSS Subscription but isn't in the regular version. On Sun, Jun 3, 2018 at 11:18 PM, Oliver <[hidden email]> wrote: Hi Brian, |
|
|
Jon,
Don't feel denigrated at all. I'm with you. I'd like to see it in the Standard Version. One of my purposes in all this was to create syntax that would include jackknife and Bayesian bootstrap and bootstrap confidence
intervals other than the three that SPSS includes, although with that said the BCa seems to be where most of the interest lies and that is one of SPSS' offerings.
Brian
From: SPSSX(r) Discussion <[hidden email]> on behalf of Jon Peck <[hidden email]>
Sent: Monday, June 4, 2018 3:07:02 PM To: [hidden email] Subject: Re: Syntax for Bootstrapping Not to denigrate Brian's work at all, but I should point out that the Bootstrapping option is included in the Standard system in SPSS Subscription. In the regular V25, the Premium version includes everything
in Professional plus Bootstrapping, Direct Marketing, Complex Samples, Neural Networks, Conjoint, and Exact Tests.
See
for details.
IMO, Bootstrapping ought to be in the Standard version, as it is for SPSS Subscription but isn't in the regular version.
On Sun, Jun 3, 2018 at 11:18 PM, Oliver
<[hidden email]> wrote:
Hi Brian, |
|
In reply to this post by David Marso
In addition to David's link/contribution, those SPSS'ers who use
MATRIX regularly would maybe dig in my functions "MATRIX - END
MATRIX functions" on
http://www.spsstools.net/en/KO-spssmacros In the description doc, search "sampling" to see the functions to do sampling with or without replacement.04.06.2018 20:17, David Marso пишет:
http://spssx-discussion.1045642.n5.nabble.com/Sampling-WITH-replacement-td5618318.html#a5620719 Perhaps useful? ===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD |
|
|
In reply to this post by bdates
Thanks much Kirill, some really good additions to one’s tool set Regards, Will WMB & Associates Statistical Services ============ mailto: [hidden email] http://home.earthlink.net/~z_statman ============ > On 6/5/2018 5:25:29 AM, Kirill Orlov ([hidden email]) wrote: |
|
|
There is some quite extensive syntax on IBM Developer Works, which produces bootstrap and jackknife results for a large number of statistics: average absolute deviation (AAD), coefficient of dispersion(COD), Minimum (MIN), Maximum (MAX), Mean (MEAN), Median (MEDIAN), median-centered coefficient of variation (MDCOV), mean-centered coefficient of variation (MNCOV), Range (RANGE), Standard deviation (STDDEV), and Variance (Var). It produces results for three approaches to confidence intervals: normal, percentile-t, and bias corrected and accelerated. It does not use the matrix mode. It's author is Mamoun Benmamoun. When I check the Developer Works site, I receive a message that the site is down temporarily. Dr. Benmamoun does leave an email to contact him about the syntax at v[hidden email].
Brian
From: SPSSX(r) Discussion <[hidden email]> on behalf of Willbaileyz @ E <[hidden email]>
Sent: Tuesday, June 5, 2018 7:12:41 AM To: [hidden email] Subject: Re: Syntax for Bootstrapping Thanks much Kirill, some really good additions to one’s tool set
Regards, Will WMB & Associates Statistical Services
============ mailto: [hidden email] http://home.earthlink.net/~z_statman ============
> On 6/5/2018 5:25:29 AM, Kirill Orlov ([hidden email]) wrote:
|
|
Administrator
|
Brian,
That might be interesting to look at. I went to IBM DW and site is up but can't locate the syntax in question. Searched Bootstrap (Several hundred hits), Mamoun (0 hits) Benmamoun (0 hits) , Bootstrap Median (0 hits) ...Any ideas? bdates wrote > There is some quite extensive syntax on IBM Developer Works, which > produces bootstrap and jackknife results for a large number of statistics: > average absolute deviation (AAD), coefficient of dispersion(COD), Minimum > (MIN), Maximum (MAX), Mean (MEAN), Median (MEDIAN), median-centered > coefficient of variation (MDCOV), mean-centered coefficient of variation > (MNCOV), Range (RANGE), Standard deviation (STDDEV), and Variance (Var). > It produces results for three approaches to confidence intervals: normal, > percentile-t, and bias corrected and accelerated. It does not use the > matrix mode. It's author is Mamoun Benmamoun. When I check the Developer > Works site, I receive a message that the site is down temporarily. Dr. > Benmamoun does leave an email to contact him about the syntax at > vmamoun@ > . > > > Brian > ________________________________ > From: SPSSX(r) Discussion < > SPSSX-L@.UGA > > on behalf of Willbaileyz @ E < > willbaileyz@ > > > Sent: Tuesday, June 5, 2018 7:12:41 AM > To: > SPSSX-L@.UGA > Subject: Re: Syntax for Bootstrapping > > > Thanks much Kirill, some really good additions to one’s tool set > > > > Regards, > > Will > > WMB & Associates > > Statistical Services > > > > ============ > > mailto: > info.statman@ > <mailto: > info.statman@ > > > > http://home.earthlink.net/~z_statman > > ============ > > > > > >> On 6/5/2018 5:25:29 AM, Kirill Orlov ( > kior@ > ) wrote: >> > In addition to David's link/contribution, those SPSS'ers who use MATRIX >> regularly would maybe dig in my functions "*MATRIX - END MATRIX >> functions*" on >> >> http://www.spsstools.net/en/KO-spssmacros >> >> In the description doc, search "sampling" to see the functions to do >> sampling with or without replacement. >> >> >> >> 04.06.2018 20:17, David Marso пишет: >> > >> http://spssx-discussion.1045642.n5.nabble.com/Sampling-WITH-replacement-td5618318.html#a5620719 >> > Perhaps useful? >> > >> > >> >> >> ===================== >> To manage your subscription to SPSSX-L, send a message to >> > LISTSERV@.UGA > (not to SPSSX-L), with no body text except the >> command. To leave the list, send the command >> SIGNOFF SPSSX-L >> For a list of commands to manage subscriptions, send the command >> INFO REFCARD >> > > ===================== To manage your subscription to SPSSX-L, send a > message to > LISTSERV@.UGA > <mailto: > LISTSERV@.UGA > > (not to SPSSX-L), with no body text except the command. To leave the > list, send the command SIGNOFF SPSSX-L For a list of commands to manage > subscriptions, send the command INFO REFCARD > > ===================== > To manage your subscription to SPSSX-L, send a message to > LISTSERV@.UGA > (not to SPSSX-L), with no body text except the > command. To leave the list, send the command > SIGNOFF SPSSX-L > For a list of commands to manage subscriptions, send the command > INFO REFCARD ----- Please reply to the list and not to my personal email. Those desiring my consulting or training services please feel free to email me. --- "Nolite dare sanctum canibus neque mittatis margaritas vestras ante porcos ne forte conculcent eas pedibus suis." Cum es damnatorum possederunt porcos iens ut salire off sanguinum cliff in abyssum?" -- Sent from: http://spssx-discussion.1045642.n5.nabble.com/ ===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD
Please reply to the list and not to my personal email.
Those desiring my consulting or training services please feel free to email me. --- "Nolite dare sanctum canibus neque mittatis margaritas vestras ante porcos ne forte conculcent eas pedibus suis." Cum es damnatorum possederunt porcos iens ut salire off sanguinum cliff in abyssum?" |
|
|
I think Brian was referring to this item on the old site. On Tue, Jun 5, 2018 at 2:23 PM, David Marso <[hidden email]> wrote: Brian, |
|
|
Jon
On the old site I found some syntax you wrote, but not the syntax to which I was referring. I'll contact Dr. Benmamoun and see if he can send it. If not, I can send what I have, maybe post to Nabble? Would that
work?
Brian
From: SPSSX(r) Discussion <[hidden email]> on behalf of Jon Peck <[hidden email]>
Sent: Tuesday, June 5, 2018 4:46:09 PM To: [hidden email] Subject: Re: Syntax for Bootstrapping I think Brian was referring to this item on the old site.
On Tue, Jun 5, 2018 at 2:23 PM, David Marso
<[hidden email]> wrote:
Brian, |
|
|
In reply to this post by Jon Peck
Hi, all. Here's a link to my Google Drive and Dr. Benmamoun's syntax. I think BootR is a revised version of BootJ with some additional solutions. There is an accompanying pdf with directions for using the syntax. In his email to me, Dr. Benmamoun indicated that some updating might be necessary since the syntax was written in 2006 in V13. Good luck.
https://drive.google.com/drive/folders/1hE-CAhgvizB-QqgQYmvh5HPnimsiESk2?usp=sharing
Brian
From: SPSSX(r) Discussion <[hidden email]> on behalf of Jon Peck <[hidden email]>
Sent: Tuesday, June 5, 2018 4:46:09 PM To: [hidden email] Subject: Re: Syntax for Bootstrapping I think Brian was referring to this item on the old site.
On Tue, Jun 5, 2018 at 2:23 PM, David Marso
<[hidden email]> wrote:
Brian, |
«
Return to SPSSX Discussion
|
1 view|%1 views
| Free forum by Nabble | Edit this page |