Re: [R] Gaussian Process Classification R packages

[Gary Black]

As a total novice and somebody lurking in the background who doesn't have a
need to use R regularly except for a one-time project, I also am surprised
by (really mainly one or two) people who answer most posts by telling people
to do their homework, spend more time studying R, etc.  Even novices know
these things, and so those replies are totally unhelpful.

For novices who have a "one time" need to use R for a project, we cannot
spend months studying as if we were going to use it every day in our jobs.
We simply read some tutorials, some books, etc. and when we run into an
issue that we cannot resolve, we ask for help.  I thought that was what
forums such as this were about - people who voluntarily answer questions for
free because they are interested in the subject.

That said, I agree totally about the need to be clear and concise when
posting questions.  But as a novice, we sometimes are not sure exactly what
the relevant information is that is needed to answer our question.

In summary, how about cutting novices some slack?  If you don't want to
answer a post in a helpful way, then just ignore it.  You don't need to
interject yourself into every post.  


-Original Message-
From: R-help [mailto:r-help-boun...@r-project.org] On Behalf Of Damjan
Krstajic
Sent: Monday, December 11, 2017 6:37 PM
To: Bert Gunter ; Jeff Newmiller

Cc: R-help ; Berry, Charles 
Subject: Re: [R] Gaussian Process Classification R packages

For the record please re-read my original message. It is clear, concise,
polite and thankful for future help. I received a reply "Google it!". Thank
you!


Thank you Jeff for your links. I am aware of them. However, they do not
point to an R package for  GP for binary classification which produces
prediction intervals.


It seems that r-help is not as it was before. Wish you all the best. Roger
and out.


From: Bert Gunter 
Sent: 12 December 2017 00:01
To: Jeff Newmiller
Cc: R-help; Damjan Krstajic; Berry, Charles
Subject: Re: [R] Gaussian Process Classification R packages

For the record:

I **was** trying to be helpful. I simply didn't know whether "I struggled"
meant that the OP had done a web search; as Chuck mentioned, when I did one,
I found what looked like possibly helpful hits. The OP's hostile response
frankly surprised me, but I see no reason to respond in kind.

Cheers,
Bert



Bert Gunter

"The trouble with having an open mind is that people keep coming along and
sticking things into it."
-- Opus (aka Berkeley Breathed in his "Bloom County" comic strip )

On Mon, Dec 11, 2017 at 3:48 PM, Jeff Newmiller
mailto:jdnew...@dcn.davis.ca.us>> wrote:
While a plea about struggling may seem appropriate to you, it is just as
content-free as a reply telling you to use Google... and like it or not,
that tit-for-tat arises due to frustration with lack of specificity as
detailed by Charles. That is, if you are constructive about documenting your
issue with a reproducible example and mentioning what you have tried and how
it failed, you won't prompt such frustrated/unhelpful responses in the
future.

Did you find [1] or [2]?

[1]
https://stats.stackexchange.com/questions/177677/gaussian-process-prediction
-interval

[2]
https://stats.stackexchange.com/questions/9131/obtaining-a-formula-for-predi
ction-limits-in-a-linear-model/9144#9144
--
Sent from my phone. Please excuse my brevity.

On December 11, 2017 9:53:01 AM PST, Damjan Krstajic
mailto:dkrsta...@hotmail.com>> wrote:
>Thank you Charles Berry for your kind reply. I don't see anything wrong 
>with the word "struggling". I have spent several hours trying various R 
>packages like kernlab and GPfit to use GP to create a binary 
>classification model which produces a prediction interval for each 
>sample. I have been struggling because with all of them you may create 
>a GP classification model but it only produces a single prediction 
>probability, and not a prediction interval of probabilities. Packages 
>that I have tried may provide a prediction interval for regression but 
>not for binary classification.
>
>
>You mention "The Gaussian Processes Web Site", have you checked how 
>many R packages are listed there?
>
>
>I have been coding in R for more than a decade and contact r-help when 
>I am struggling (I don't see anything wrong with this word) to find a 
>solution in R. Replies like "Google it!" are below my level of my 
>communication and understanding of others.
>
>
>Best wishes
>
>DK
>
>
>
>From: Berry, Charles mailto:ccbe...@ucsd.edu>>
>Sent: 11 December 2017 17:04
>To: Damjan Krstajic
>Cc: Bert Gunter; r-help@r-project.org
>Subject: Re: Gaussian Process Classification R packages
>
>
>> On Dec 11, 2017, at 8:06 AM, Damjan Krstajic 
>> mailto:dkrsta...@hotmail.com>>
>wrote:
>>
>> I have kindly asked for help and I am sad to receive such a reply
>from some on the r-help list.
>>
>>
>
>Well, you only said you were `struggling' to find 

Re: [R] Gaussian Process Classification R packages

[Bert Gunter]

For the record:

I **was** trying to be helpful. I simply didn't know whether "I struggled"
meant that the OP had done a web search; as Chuck mentioned, when I did
one, I found what looked like possibly helpful hits. The OP's hostile
response frankly surprised me, but I see no reason to respond in kind.

Cheers,
Bert



Bert Gunter

"The trouble with having an open mind is that people keep coming along and
sticking things into it."
-- Opus (aka Berkeley Breathed in his "Bloom County" comic strip )

On Mon, Dec 11, 2017 at 3:48 PM, Jeff Newmiller 
wrote:

> While a plea about struggling may seem appropriate to you, it is just as
> content-free as a reply telling you to use Google... and like it or not,
> that tit-for-tat arises due to frustration with lack of specificity as
> detailed by Charles. That is, if you are constructive about documenting
> your issue with a reproducible example and mentioning what you have tried
> and how it failed, you won't prompt such frustrated/unhelpful responses in
> the future.
>
> Did you find [1] or [2]?
>
> [1] https://stats.stackexchange.com/questions/177677/gaussian-
> process-prediction-interval
>
> [2] https://stats.stackexchange.com/questions/9131/obtaining-
> a-formula-for-prediction-limits-in-a-linear-model/9144#9144
> --
> Sent from my phone. Please excuse my brevity.
>
> On December 11, 2017 9:53:01 AM PST, Damjan Krstajic <
> dkrsta...@hotmail.com> wrote:
> >Thank you Charles Berry for your kind reply. I don't see anything wrong
> >with the word "struggling". I have spent several hours trying various R
> >packages like kernlab and GPfit to use GP to create a binary
> >classification model which produces a prediction interval for each
> >sample. I have been struggling because with all of them you may create
> >a GP classification model but it only produces a single prediction
> >probability, and not a prediction interval of probabilities. Packages
> >that I have tried may provide a prediction interval for regression but
> >not for binary classification.
> >
> >
> >You mention "The Gaussian Processes Web Site", have you checked how
> >many R packages are listed there?
> >
> >
> >I have been coding in R for more than a decade and contact r-help when
> >I am struggling (I don't see anything wrong with this word) to find a
> >solution in R. Replies like "Google it!" are below my level of my
> >communication and understanding of others.
> >
> >
> >Best wishes
> >
> >DK
> >
> >
> >
> >From: Berry, Charles 
> >Sent: 11 December 2017 17:04
> >To: Damjan Krstajic
> >Cc: Bert Gunter; r-help@r-project.org
> >Subject: Re: Gaussian Process Classification R packages
> >
> >
> >> On Dec 11, 2017, at 8:06 AM, Damjan Krstajic 
> >wrote:
> >>
> >> I have kindly asked for help and I am sad to receive such a reply
> >from some on the r-help list.
> >>
> >>
> >
> >Well, you only said you were `struggling' to find a package.
> >
> >Bert may well have done the Google search himself and found numerous
> >resources on such models including links to R (as I did, see below).
> >If so, his response seems quite natural.
> >
> >Perhaps, you need to say what is wrong with the hits you got and the
> >packages that they describe to keep a potential response from running
> >in the wrong direction. Perhaps, you have misunderstood the
> >capabilities of a package or failed to grasp an inobvious way to use
> >the package to reach your goal.  In any case, providing some background
> >of why you think the obvious leads do not work in your case can be
> >helpful.
> >
> >Doing that search myself I see links to R packages, R functions, and to
> >"The Gaussian Processes Web Site" which has a table of possibly
> >relevant softwares.  It seems like there is a lot there to digest.
> >
> >HTH,
> >
> >Chuck
> >
> >> I did google it prior to sending my request, and I could not find any
> >R package which provides GP classification model which produces
> >prediction intervals for each sample. I would be grateful if anybody
> >could inform me about it. Thank you.
> >>
> >>
> >> 
> >> From: Bert Gunter 
> >> Sent: 11 December 2017 15:50
> >> To: Damjan Krstajic
> >> Cc: r-help@r-project.org
> >> Subject: Re: [R] Gaussian Process Classification R packages
> >>
> >> Google it!
> >>
> >> "R Gaussian process model binary classification."
> >>
> >> Cheers,
> >> Bert
> >>
> >
> >
> >
> >   [[alternative HTML version deleted]]
> >
> >__
> >R-help@r-project.org mailing list -- To UNSUBSCRIBE and more, see
> >https://stat.ethz.ch/mailman/listinfo/r-help
> >PLEASE do read the posting guide
> >http://www.R-project.org/posting-guide.html
> >and provide commented, minimal, self-contained, reproducible code.
>
> __
> R-help@r-project.org mailing list -- To UNSUBSCRIBE and more, see
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide http://www.R

Re: [R] Gaussian Process Classification R packages

[Jeff Newmiller]

While a plea about struggling may seem appropriate to you, it is just as 
content-free as a reply telling you to use Google... and like it or not, that 
tit-for-tat arises due to frustration with lack of specificity as detailed by 
Charles. That is, if you are constructive about documenting your issue with a 
reproducible example and mentioning what you have tried and how it failed, you 
won't prompt such frustrated/unhelpful responses in the future.

Did you find [1] or [2]?

[1] 
https://stats.stackexchange.com/questions/177677/gaussian-process-prediction-interval

[2] 
https://stats.stackexchange.com/questions/9131/obtaining-a-formula-for-prediction-limits-in-a-linear-model/9144#9144
-- 
Sent from my phone. Please excuse my brevity.

On December 11, 2017 9:53:01 AM PST, Damjan Krstajic  
wrote:
>Thank you Charles Berry for your kind reply. I don't see anything wrong
>with the word "struggling". I have spent several hours trying various R
>packages like kernlab and GPfit to use GP to create a binary
>classification model which produces a prediction interval for each
>sample. I have been struggling because with all of them you may create
>a GP classification model but it only produces a single prediction
>probability, and not a prediction interval of probabilities. Packages
>that I have tried may provide a prediction interval for regression but
>not for binary classification.
>
>
>You mention "The Gaussian Processes Web Site", have you checked how
>many R packages are listed there?
>
>
>I have been coding in R for more than a decade and contact r-help when
>I am struggling (I don't see anything wrong with this word) to find a
>solution in R. Replies like "Google it!" are below my level of my
>communication and understanding of others.
>
>
>Best wishes
>
>DK
>
>
>
>From: Berry, Charles 
>Sent: 11 December 2017 17:04
>To: Damjan Krstajic
>Cc: Bert Gunter; r-help@r-project.org
>Subject: Re: Gaussian Process Classification R packages
>
>
>> On Dec 11, 2017, at 8:06 AM, Damjan Krstajic 
>wrote:
>>
>> I have kindly asked for help and I am sad to receive such a reply
>from some on the r-help list.
>>
>>
>
>Well, you only said you were `struggling' to find a package.
>
>Bert may well have done the Google search himself and found numerous
>resources on such models including links to R (as I did, see below). 
>If so, his response seems quite natural.
>
>Perhaps, you need to say what is wrong with the hits you got and the
>packages that they describe to keep a potential response from running
>in the wrong direction. Perhaps, you have misunderstood the
>capabilities of a package or failed to grasp an inobvious way to use
>the package to reach your goal.  In any case, providing some background
>of why you think the obvious leads do not work in your case can be
>helpful.
>
>Doing that search myself I see links to R packages, R functions, and to
>"The Gaussian Processes Web Site" which has a table of possibly
>relevant softwares.  It seems like there is a lot there to digest.
>
>HTH,
>
>Chuck
>
>> I did google it prior to sending my request, and I could not find any
>R package which provides GP classification model which produces
>prediction intervals for each sample. I would be grateful if anybody
>could inform me about it. Thank you.
>>
>>
>> 
>> From: Bert Gunter 
>> Sent: 11 December 2017 15:50
>> To: Damjan Krstajic
>> Cc: r-help@r-project.org
>> Subject: Re: [R] Gaussian Process Classification R packages
>>
>> Google it!
>>
>> "R Gaussian process model binary classification."
>>
>> Cheers,
>> Bert
>>
>
>
>
>   [[alternative HTML version deleted]]
>
>__
>R-help@r-project.org mailing list -- To UNSUBSCRIBE and more, see
>https://stat.ethz.ch/mailman/listinfo/r-help
>PLEASE do read the posting guide
>http://www.R-project.org/posting-guide.html
>and provide commented, minimal, self-contained, reproducible code.

__
R-help@r-project.org mailing list -- To UNSUBSCRIBE and more, see
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.

Re: [R] OT -- isotonic regression subject to bound constraints.

[Rolf Turner]

On 12/12/17 07:15, Rolf Turner wrote:


Well, I could argue that it's not *completely* OT since my question is 
motivated by an enquiry that I received in respect of a CRAN package 
"Iso" that I wrote and maintain.


The question is this:  Given observations y_1, ..., y_n, what is the 
solution to the problem:


   minimise \sum_{i=1}^n (y_i - y_i^*)^2

with respect to y_1^*, ..., y_n^* subject to the "isotonic" constraint
y_1^* <= y_2^* <= ... <= y_n^* and the *additional8 bound constraint
a <= y_1^* and y_n^* <= b, where a and b are given constants?




Scrub that question!  *Just* after I sent it (wouldn't you know!) I got 
an email from my original enquirer telling me that he'd found the 
solution in the package OrdMonReg on CRAN.


Sorry for the noise.

cheers,

Rolf Turner


--
Technical Editor ANZJS
Department of Statistics
University of Auckland
Phone: +64-9-373-7599 ext. 88276

__
R-help@r-project.org mailing list -- To UNSUBSCRIBE and more, see
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.

[R] OT -- isotonic regression subject to bound constraints.

[Rolf Turner]

Well, I could argue that it's not *completely* OT since my question is 
motivated by an enquiry that I received in respect of a CRAN package 
"Iso" that I wrote and maintain.


The question is this:  Given observations y_1, ..., y_n, what is the 
solution to the problem:


  minimise \sum_{i=1}^n (y_i - y_i^*)^2

with respect to y_1^*, ..., y_n^* subject to the "isotonic" constraint
y_1^* <= y_2^* <= ... <= y_n^* and the *additional8 bound constraint
a <= y_1^* and y_n^* <= b, where a and b are given constants?

I have googled around a bit (unsuccessfully) and have asked this 
question on crossvalidated a couple of days ago, with no response whatever.


So I thought that I might try the super-knowledgeable R community, in 
the hope that someone out there might be able to tell me something useful.


Note that the question can be expressed as finding the projection of the 
point (y_1, ..., y_n) onto the intersection of the isotonic cone and

the hypercube [a,b]^n.

At first I thought that protecting onto the isotonic cone and then 
projection that result onto the hypercube might work, but I am now 
pretty sure that is hopelessly naive.


Any hints?  Ta.

cheers,

Rolf Turner


--
Technical Editor ANZJS
Department of Statistics
University of Auckland
Phone: +64-9-373-7599 ext. 88276

__
R-help@r-project.org mailing list -- To UNSUBSCRIBE and more, see
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.

Re: [R] Data cleaning & Data preparation, what do R users want?

[Robert Wilkins]

Dominik (and others)

If it is indeed still the biggest paint point, even in 2017, then maybe we
can do something about that, with more efforts at different user interface
design and try-outs with them on specialized datasets.
[ The fact that in some specialties, such as clinical trials, for example,
getting access to public domain datasets (and not having to use a tiny
"toy" dataset, which nobody will pay attention to, does make it harder].

It would help if academia (both comp-sci and statistics departments) would
support those who invest resources in drafting and test-driving new product
designs. If, in the year 2017, it is still a big pain point, doesn't that
make sense. More speculative work in statistical programming language
design has not been a priority in academia since before 1980.

On Thu, Nov 30, 2017 at 4:11 AM, Dominik Schneider <
dominik.schnei...@colorado.edu> wrote:

> I would agree that getting data into R from various sources is the biggest
> pain point. Even if there is an api, the results are not always consistent
> and you have to do lots of dimension checking to get it right. Or there
> isn't an open api at all and you have to hack it by web scraping or
> otherwise- http://enpiar.com/2017/08/11/one-hour-package/
>
> On Thu, Nov 30, 2017 at 1:00 AM, Jim Lemon  wrote:
>
>> Hi again,
>> Typo in the last email. Should read "about 40 standard deviations".
>>
>> Jim
>>
>> On Thu, Nov 30, 2017 at 10:54 AM, Jim Lemon  wrote:
>> > Hi Robert,
>> > People want different levels of automation in the software they use.
>> > What concerns many of us is the desire for the function
>> > "figure-out-what-this-data-is-import-it-and-get-rid-of-bad-values".
>> > Such users typically want something that justifies its use by being
>> > written by someone who seems to know what they're doing and lots of
>> > other people use it. One advantage of many R functions is their
>> > modular construction. This encourages users to at least consider the
>> > steps that are taken rather than just accept what comes out of that
>> > long tube.
>> >
>> > Take the contentious problem of outlier identification. If I just let
>> > the black box peel off some values, I don't know what I have lost. On
>> > the other hand, if I import data and examine it with a summary
>> > function, I may find that one woman has a height of 5.2 meters. I can
>> > range check by looking up the Guinness Book of Records. It's an
>> > outlier. I can estimate the probability of such a height.  Hmm, about
>> > 4 standard deviations above the mean. It's an outlier. I can attempt a
>> > Sherlock Holmes. "Watson, I conclude that an imperial measure (5'2")
>> > has been recorded as a metric value". It's not an outlier.
>> >
>> > The more R gravitates toward "black box" functions, the more some
>> > users are encouraged to let them do the work.You pays your money and
>> > you takes your chances.
>> >
>> > Jim
>> >
>> >
>> > On Thu, Nov 30, 2017 at 3:37 AM, Robert Wilkins 
>> wrote:
>> >> R has a very wide audience, clinical research, astronomy, psychology,
>> and
>> >> so on and so on.
>> >> I would consider data analysis work to be three stages: data
>> preparation,
>> >> statistical analysis, and producing the report.
>> >> This regards the process of getting the data ready for analysis and
>> >> reporting, sometimes called "data cleaning" or "data munging" or "data
>> >> wrangling".
>> >>
>> >> So as regards tools for data preparation, speaking to the highly
>> diverse
>> >> audience mentioned, here is my question:
>> >>
>> >> What do you want?
>> >> Or are you already quite happy with the range of tools that is
>> currently
>> >> before you?
>> >>
>> >> [BTW,  I posed the same question last week to the r-devel list, and was
>> >> advised that r-help might be a more suitable audience by one of the
>> >> moderators.]
>> >>
>> >> Robert Wilkins
>> >>
>> >> [[alternative HTML version deleted]]
>> >>
>> >> __
>> >> R-help@r-project.org mailing list -- To UNSUBSCRIBE and more, see
>> >> https://stat.ethz.ch/mailman/listinfo/r-help
>> >> PLEASE do read the posting guide http://www.R-project.org/posti
>> ng-guide.html
>> >> and provide commented, minimal, self-contained, reproducible code.
>>
>> __
>> R-help@r-project.org mailing list -- To UNSUBSCRIBE and more, see
>> https://stat.ethz.ch/mailman/listinfo/r-help
>> PLEASE do read the posting guide http://www.R-project.org/posti
>> ng-guide.html
>> and provide commented, minimal, self-contained, reproducible code.
>>
>
>

[[alternative HTML version deleted]]

__
R-help@r-project.org mailing list -- To UNSUBSCRIBE and more, see
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.

Re: [R] Gaussian Process Classification R packages

[Berry, Charles]

> On Dec 11, 2017, at 8:06 AM, Damjan Krstajic  wrote:
> 
> I have kindly asked for help and I am sad to receive such a reply from some 
> on the r-help list.
> 
> 

Well, you only said you were `struggling' to find a package.

Bert may well have done the Google search himself and found numerous resources 
on such models including links to R (as I did, see below).  If so, his response 
seems quite natural.

Perhaps, you need to say what is wrong with the hits you got and the packages 
that they describe to keep a potential response from running in the wrong 
direction. Perhaps, you have misunderstood the capabilities of a package or 
failed to grasp an inobvious way to use the package to reach your goal.  In any 
case, providing some background of why you think the obvious leads do not work 
in your case can be helpful.

Doing that search myself I see links to R packages, R functions, and to "The 
Gaussian Processes Web Site" which has a table of possibly relevant softwares.  
It seems like there is a lot there to digest.

HTH,

Chuck

> I did google it prior to sending my request, and I could not find any R 
> package which provides GP classification model which produces prediction 
> intervals for each sample. I would be grateful if anybody could inform me 
> about it. Thank you.
> 
> 
> 
> From: Bert Gunter 
> Sent: 11 December 2017 15:50
> To: Damjan Krstajic
> Cc: r-help@r-project.org
> Subject: Re: [R] Gaussian Process Classification R packages
> 
> Google it!
> 
> "R Gaussian process model binary classification."
> 
> Cheers,
> Bert
> 

__
R-help@r-project.org mailing list -- To UNSUBSCRIBE and more, see
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.

Re: [R] Gaussian Process Classification R packages

[Bert Gunter]

Google it!

"R Gaussian process model binary classification."

Cheers,
Bert



Bert Gunter

"The trouble with having an open mind is that people keep coming along and
sticking things into it."
-- Opus (aka Berkeley Breathed in his "Bloom County" comic strip )

On Mon, Dec 11, 2017 at 4:53 AM, Damjan Krstajic 
wrote:

> Dear All,
>
>
> I am struggling to find an R package which contains a function for
> building a Gaussian Process model for binary classification which may
> produce prediction intervals for predicted probabilities. I would be
> grateful if somebody could point me to such package.
>
>
> Thank you very much in advance.
>
> DK
>
> [[alternative HTML version deleted]]
>
> __
> R-help@r-project.org mailing list -- To UNSUBSCRIBE and more, see
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide http://www.R-project.org/
> posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
>

[[alternative HTML version deleted]]

__
R-help@r-project.org mailing list -- To UNSUBSCRIBE and more, see
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.