{ "id": 9493, "title": "please tell me about \"quantile regression\"", "url": "/competitions/acquire-valued-shoppers-challenge/discussion/9493", "author_name": "", "post_date": "2014-06-19T06:19:24.437Z", "votes": null, "comment_count": 9, "views": 4747, "content": "

hello all,

\n\n

i am a japanese student studying statistics

\n

i read this dashboard and knew "quantile regression"

\n\n

how can i study quantile regression?

\n

how can i conduct quantile regression?

\n

and is there any R package of quantile regression?

", "messages": [ { "id": "49288", "postDate": "06/19/2014 06:19:24", "content": "

hello all,

\n\n

i am a japanese student studying statistics

\n

i read this dashboard and knew "quantile regression"

\n\n

how can i study quantile regression?

\n

how can i conduct quantile regression?

\n

and is there any R package of quantile regression?

", "rawMarkdown": "", "votes": null }, { "id": "49305", "postDate": "06/19/2014 15:13:28", "content": "

I am not an expert, but I can answer a couple of your questions.  :-)  quantreg is an R package that implements quantile regression, as well as a number of helper functions.

\n\n

Vowpal Wabbit (as noted in Triskelion's tutorial on this forum) implements quantile regression as well, with some added tricks and niceties.

", "rawMarkdown": "", "votes": null }, { "id": "49323", "postDate": "06/20/2014 00:36:29", "content": "

 

\n

Just for the record, I'm going to leave this here:

\n

GradientBoostingRegressor class of Scikit-learn in Python implements Quantile Regression as well.

", "rawMarkdown": "", "votes": null }, { "id": "49327", "postDate": "06/20/2014 03:35:38", "content": "

In Japanese, quantile regression is "分位点回帰" or sometimes "区分点回帰." If you search this word, you can get some lecture notes (of course written in Japanese) which is useful to understand its concepts. But there are few Japanese documents on the calculation/programming.

", "rawMarkdown": "", "votes": null }, { "id": "49374", "postDate": "06/22/2014 07:24:21", "content": "

FYI, besides the mentioned quantreg package in R, there are quantregForest package using random forest for quantile regression, and also gbm (Gradient Boosting Machine) package with quantile regression.

", "rawMarkdown": "", "votes": null }, { "id": "49456", "postDate": "06/24/2014 05:38:21", "content": "

i counducted R code and got some results,

\n

but i was not able to interpret them...

\n\n

could you tell me how to read and interpret the results??

\n\n

i used R package of "quantregForest"

\n

but it did not work

\n

does it work even if the response variance is binary??

", "rawMarkdown": "", "votes": null }, { "id": "49682", "postDate": "06/27/2014 07:19:35", "content": "

Hi, yr~

\n

I try to search some documents about the Quantile Regression in R, and find some,

\n

But I cannot understand how to use the quantreg as it was used in Vowpal Wabbit.

\n\n

If it is ok, I am afraid that can you give us some examples?

\n

Thanks a lot in advance!

", "rawMarkdown": "", "votes": null }, { "id": "49699", "postDate": "06/27/2014 12:56:06", "content": "

Hi all,

\n\n

@hideoJapan, it seems quantregForest doesn't work on binary response. Give GBM a try or you might also try to regress to the repeattrips variable.

\n\n

@wadefighting, Pls find in the attachment a simple demo of quantreg, GBM (with quantile loss), and quantregForest. It uses a toy dataset, but you can easily work on your own dataset in a similar manner.

", "rawMarkdown": "", "votes": null }, { "id": "49759", "postDate": "06/28/2014 07:59:52", "content": "

Many Thanks for yr!!!

", "rawMarkdown": "", "votes": null }, { "id": "49939", "postDate": "07/02/2014 06:40:11", "content": "

thanks a lot!!!

\n

i will try!!

", "rawMarkdown": "", "votes": null } ], "comments": [ { "id": 49305, "author_name": "philculliton", "author_url": "", "post_date": "06/19/2014 15:13:28", "content": "

I am not an expert, but I can answer a couple of your questions.  :-)  quantreg is an R package that implements quantile regression, as well as a number of helper functions.

\n\n

Vowpal Wabbit (as noted in Triskelion's tutorial on this forum) implements quantile regression as well, with some added tricks and niceties.

", "votes": null, "replies": [] }, { "id": 49323, "author_name": "rohits", "author_url": "", "post_date": "06/20/2014 00:36:29", "content": "

 

\n

Just for the record, I'm going to leave this here:

\n

GradientBoostingRegressor class of Scikit-learn in Python implements Quantile Regression as well.

", "votes": null, "replies": [] }, { "id": 49327, "author_name": "ktgrs197680", "author_url": "", "post_date": "06/20/2014 03:35:38", "content": "

In Japanese, quantile regression is "分位点回帰" or sometimes "区分点回帰." If you search this word, you can get some lecture notes (of course written in Japanese) which is useful to understand its concepts. But there are few Japanese documents on the calculation/programming.

", "votes": null, "replies": [] }, { "id": 49374, "author_name": "chenglongchen", "author_url": "", "post_date": "06/22/2014 07:24:21", "content": "

FYI, besides the mentioned quantreg package in R, there are quantregForest package using random forest for quantile regression, and also gbm (Gradient Boosting Machine) package with quantile regression.

", "votes": null, "replies": [] }, { "id": 49456, "author_name": "user1190597", "author_url": "", "post_date": "06/24/2014 05:38:21", "content": "

i counducted R code and got some results,

\n

but i was not able to interpret them...

\n\n

could you tell me how to read and interpret the results??

\n\n

i used R package of "quantregForest"

\n

but it did not work

\n

does it work even if the response variance is binary??

", "votes": null, "replies": [] }, { "id": 49682, "author_name": "wadefighting", "author_url": "", "post_date": "06/27/2014 07:19:35", "content": "

Hi, yr~

\n

I try to search some documents about the Quantile Regression in R, and find some,

\n

But I cannot understand how to use the quantreg as it was used in Vowpal Wabbit.

\n\n

If it is ok, I am afraid that can you give us some examples?

\n

Thanks a lot in advance!

", "votes": null, "replies": [] }, { "id": 49699, "author_name": "chenglongchen", "author_url": "", "post_date": "06/27/2014 12:56:06", "content": "

Hi all,

\n\n

@hideoJapan, it seems quantregForest doesn't work on binary response. Give GBM a try or you might also try to regress to the repeattrips variable.

\n\n

@wadefighting, Pls find in the attachment a simple demo of quantreg, GBM (with quantile loss), and quantregForest. It uses a toy dataset, but you can easily work on your own dataset in a similar manner.

", "votes": null, "replies": [] }, { "id": 49759, "author_name": "wadefighting", "author_url": "", "post_date": "06/28/2014 07:59:52", "content": "

Many Thanks for yr!!!

", "votes": null, "replies": [] }, { "id": 49939, "author_name": "user1190597", "author_url": "", "post_date": "07/02/2014 06:40:11", "content": "

thanks a lot!!!

\n

i will try!!

", "votes": null, "replies": [] } ], "raw_markdown_by_id": { "49288": "", "49305": "", "49323": "", "49327": "", "49374": "", "49456": "", "49682": "", "49699": "", "49759": "", "49939": "" }, "source": "meta" }