I use mainly R in competitions.\nIf I merge team with someone who uses Python and our hopefully awesome solution contains both Python and R.\nIs the submission eligible for the R Prize?
\n\nOr must the submission solely use R?
\n\nThanks!
", "messages": [ { "id": "387810", "postDate": "09/15/2018 17:53:01", "content": "I use mainly R in competitions.\nIf I merge team with someone who uses Python and our hopefully awesome solution contains both Python and R.\nIs the submission eligible for the R Prize?
\n\nOr must the submission solely use R?
\n\nThanks!
", "rawMarkdown": "I use mainly R in competitions.\nIf I merge team with someone who uses Python and our hopefully awesome solution contains both Python and R.\nIs the submission eligible for the R Prize?\n\nOr must the submission solely use R?\n\nThanks!", "votes": null }, { "id": "388935", "postDate": "09/17/2018 20:56:54", "content": "The intent of this prize is to showcase the use of R in your solution. In the spirit of that intent, you should be able to self-assess whether your submission is an honorable use of R. And it will be at the host's discretion to measure that usage.
\n\nThat said, it is acceptable to use some python with the reticulate package for the python-R translation. However, in an extreme example, if your entire model is written in Python, and then you were to call one line of Reticulate to do the final translation to R, then the host would feel obliged to disqualify this entry.
", "rawMarkdown": "The intent of this prize is to showcase the use of R in your solution. In the spirit of that intent, you should be able to self-assess whether your submission is an honorable use of R. And it will be at the host's discretion to measure that usage. \n\nThat said, it is acceptable to use *some* python with the reticulate package for the python-R translation. However, in an extreme example, if your *entire* model is written in Python, and then you were to call one line of Reticulate to do the final translation to R, then the host would feel obliged to disqualify this entry.", "votes": null }, { "id": "388953", "postDate": "09/17/2018 22:18:09", "content": "hello Julia, if i have a model in R that could score high, but i did not pick it in the end. can i still submit it for R prize or it has to be one of the final 2 submissions?
", "rawMarkdown": "hello Julia, if i have a model in R that could score high, but i did not pick it in the end. can i still submit it for R prize or it has to be one of the final 2 submissions?", "votes": null }, { "id": "388958", "postDate": "09/17/2018 22:30:41", "content": "@Shubin It has to be one of the final 2 submissions.
", "rawMarkdown": "Shubin It has to be one of the final 2 submissions.", "votes": null }, { "id": "388987", "postDate": "09/18/2018 00:02:49", "content": "@Julia, thanks for explaining. I usually use both R and Python in my normal work (R for eda, feature engineering, and simple ensembling; Python for any kind of modeling). Could you please take the time to address some questions that I have?
\n\nThanks!
", "rawMarkdown": "Julia, thanks for explaining. I usually use both R and Python in my normal work (R for eda, feature engineering, and simple ensembling; Python for any kind of modeling). Could you please take the time to address some questions that I have?\n\n 1. *\"you should be able to self-assess whether your submission is an honorable use of R\".* This is still a bit unclear to me. I don't really know what qualifies as \"honorable\". Could you please provide a little more detail on what it means to be \"honorable\"? \n 2. *\"it is acceptable to use some python with the reticulate package for the python-R translation\"*. Does this mean that if Python is used, then it must be run via the reticulate package? Or can I do my feature engineering in R and save the features to disk, then load the features into python to build my models?\n\nThanks!", "votes": null }, { "id": "389154", "postDate": "09/18/2018 08:57:49", "content": "Hi @Matt Motoki,
\n\nI'm not sure about these issues just like you,\nFrom what i understood, there's no black and white, and the organizers will decide when the time comes, what feels \"honorable\" enough.
\n\nFrom what I think, doing simple feature engineering or processing in R and doing the modelling in Python isn't enough.\nIf R is sponsoring this competition, they want to show that R is as powerful as an ML tool just as Python, and want Data Scientists to use it more than they do now. (As an R user myself, i feel the same way when i see most kernels on kaggle being in Python).\nI think the R prize will be legible for teams that used R as a significant ML tool in the competition, and teams won't be able to \"hack\" their way into the R prize by simply pushing in some processing in R.\nIt will have to be really significant part of the solution.
", "rawMarkdown": "Hi @Matt Motoki,\n\nI'm not sure about these issues just like you,\nFrom what i understood, there's no black and white, and the organizers will decide when the time comes, what feels \"honorable\" enough.\n\nFrom what I think, doing simple feature engineering or processing in R and doing the modelling in Python isn't enough.\nIf R is sponsoring this competition, they want to show that R is as powerful as an ML tool just as Python, and want Data Scientists to use it more than they do now. (As an R user myself, i feel the same way when i see most kernels on kaggle being in Python).\nI think the R prize will be legible for teams that used R as a significant ML tool in the competition, and teams won't be able to \"hack\" their way into the R prize by simply pushing in some processing in R.\nIt will have to be really significant part of the solution.", "votes": null }, { "id": "389363", "postDate": "09/18/2018 15:38:57", "content": "AmirH is spot on. The special prize’s award will be at the host’s discretion based on the principle of whether you are demonstrating you can win this competition using the best of R, a rich language which offers extensive possibility and inventiveness. So while you may employ some interoperability capabilities of the language’s design, you should expect the host to evaluate the award based on whether R’s use is part of the entire model workflow. You should not expect a response with some quantitative measure of exactly how much of a non-R language you may use. The point is whether you have made a submission with this principle and the prize’s intent in mind.
", "rawMarkdown": "AmirH is spot on. The special prize’s award will be at the host’s discretion based on the principle of whether you are demonstrating you can win this competition using the best of R, a rich language which offers extensive possibility and inventiveness. So while you may employ some interoperability capabilities of the language’s design, you should expect the host to evaluate the award based on whether R’s use is part of the entire model workflow. You should not expect a response with some quantitative measure of exactly how much of a non-R language you may use. The point is whether you have made a submission with this principle and the prize’s intent in mind.", "votes": null }, { "id": "389520", "postDate": "09/18/2018 21:14:19", "content": "@AmirH, @Julia Elliott Got it! Thank you both for the responses, I think I understand the situation better now.
", "rawMarkdown": "AmirH, @Julia Elliott Got it! Thank you both for the responses, I think I understand the situation better now.", "votes": null }, { "id": "391299", "postDate": "09/21/2018 14:59:19", "content": "@julia
\n\nIs the use of external executables such as libffm as part of the solution is acceptable for the R Prize?
If one of the submission is done is R and the other is not, would the R submission alone qualify for the R Prize?
Thanks for the clarification!
", "rawMarkdown": "julia \n\n1. Is the use of external executables such as libffm as part of the solution is acceptable for the R Prize?\n\n2. If one of the submission is done is R and the other is not, would the R submission alone qualify for the R Prize?\n\nThanks for the clarification!", "votes": null } ], "comments": [ { "id": 388935, "author_name": "juliaelliott", "author_url": "", "post_date": "09/17/2018 20:56:54", "content": "The intent of this prize is to showcase the use of R in your solution. In the spirit of that intent, you should be able to self-assess whether your submission is an honorable use of R. And it will be at the host's discretion to measure that usage.
\n\nThat said, it is acceptable to use some python with the reticulate package for the python-R translation. However, in an extreme example, if your entire model is written in Python, and then you were to call one line of Reticulate to do the final translation to R, then the host would feel obliged to disqualify this entry.
", "votes": null, "replies": [ { "id": 388953, "author_name": "lishubin", "author_url": "", "post_date": "09/17/2018 22:18:09", "content": "hello Julia, if i have a model in R that could score high, but i did not pick it in the end. can i still submit it for R prize or it has to be one of the final 2 submissions?
", "votes": null, "replies": [] }, { "id": 388958, "author_name": "juliaelliott", "author_url": "", "post_date": "09/17/2018 22:30:41", "content": "@Shubin It has to be one of the final 2 submissions.
", "votes": null, "replies": [] }, { "id": 388987, "author_name": "mmotoki", "author_url": "", "post_date": "09/18/2018 00:02:49", "content": "@Julia, thanks for explaining. I usually use both R and Python in my normal work (R for eda, feature engineering, and simple ensembling; Python for any kind of modeling). Could you please take the time to address some questions that I have?
\n\nThanks!
", "votes": null, "replies": [] }, { "id": 389154, "author_name": "tpthegreat", "author_url": "", "post_date": "09/18/2018 08:57:49", "content": "Hi @Matt Motoki,
\n\nI'm not sure about these issues just like you,\nFrom what i understood, there's no black and white, and the organizers will decide when the time comes, what feels \"honorable\" enough.
\n\nFrom what I think, doing simple feature engineering or processing in R and doing the modelling in Python isn't enough.\nIf R is sponsoring this competition, they want to show that R is as powerful as an ML tool just as Python, and want Data Scientists to use it more than they do now. (As an R user myself, i feel the same way when i see most kernels on kaggle being in Python).\nI think the R prize will be legible for teams that used R as a significant ML tool in the competition, and teams won't be able to \"hack\" their way into the R prize by simply pushing in some processing in R.\nIt will have to be really significant part of the solution.
", "votes": null, "replies": [] }, { "id": 389363, "author_name": "juliaelliott", "author_url": "", "post_date": "09/18/2018 15:38:57", "content": "AmirH is spot on. The special prize’s award will be at the host’s discretion based on the principle of whether you are demonstrating you can win this competition using the best of R, a rich language which offers extensive possibility and inventiveness. So while you may employ some interoperability capabilities of the language’s design, you should expect the host to evaluate the award based on whether R’s use is part of the entire model workflow. You should not expect a response with some quantitative measure of exactly how much of a non-R language you may use. The point is whether you have made a submission with this principle and the prize’s intent in mind.
", "votes": null, "replies": [] }, { "id": 389520, "author_name": "mmotoki", "author_url": "", "post_date": "09/18/2018 21:14:19", "content": "@AmirH, @Julia Elliott Got it! Thank you both for the responses, I think I understand the situation better now.
", "votes": null, "replies": [] } ] }, { "id": 391299, "author_name": "yilihome", "author_url": "", "post_date": "09/21/2018 14:59:19", "content": "@julia
\n\nIs the use of external executables such as libffm as part of the solution is acceptable for the R Prize?
If one of the submission is done is R and the other is not, would the R submission alone qualify for the R Prize?
Thanks for the clarification!
", "votes": null, "replies": [] } ], "raw_markdown_by_id": { "387810": "I use mainly R in competitions.\nIf I merge team with someone who uses Python and our hopefully awesome solution contains both Python and R.\nIs the submission eligible for the R Prize?\n\nOr must the submission solely use R?\n\nThanks!", "388935": "The intent of this prize is to showcase the use of R in your solution. In the spirit of that intent, you should be able to self-assess whether your submission is an honorable use of R. And it will be at the host's discretion to measure that usage. \n\nThat said, it is acceptable to use *some* python with the reticulate package for the python-R translation. However, in an extreme example, if your *entire* model is written in Python, and then you were to call one line of Reticulate to do the final translation to R, then the host would feel obliged to disqualify this entry.", "388953": "hello Julia, if i have a model in R that could score high, but i did not pick it in the end. can i still submit it for R prize or it has to be one of the final 2 submissions?", "388958": "Shubin It has to be one of the final 2 submissions.", "388987": "Julia, thanks for explaining. I usually use both R and Python in my normal work (R for eda, feature engineering, and simple ensembling; Python for any kind of modeling). Could you please take the time to address some questions that I have?\n\n 1. *\"you should be able to self-assess whether your submission is an honorable use of R\".* This is still a bit unclear to me. I don't really know what qualifies as \"honorable\". Could you please provide a little more detail on what it means to be \"honorable\"? \n 2. *\"it is acceptable to use some python with the reticulate package for the python-R translation\"*. Does this mean that if Python is used, then it must be run via the reticulate package? Or can I do my feature engineering in R and save the features to disk, then load the features into python to build my models?\n\nThanks!", "389154": "Hi @Matt Motoki,\n\nI'm not sure about these issues just like you,\nFrom what i understood, there's no black and white, and the organizers will decide when the time comes, what feels \"honorable\" enough.\n\nFrom what I think, doing simple feature engineering or processing in R and doing the modelling in Python isn't enough.\nIf R is sponsoring this competition, they want to show that R is as powerful as an ML tool just as Python, and want Data Scientists to use it more than they do now. (As an R user myself, i feel the same way when i see most kernels on kaggle being in Python).\nI think the R prize will be legible for teams that used R as a significant ML tool in the competition, and teams won't be able to \"hack\" their way into the R prize by simply pushing in some processing in R.\nIt will have to be really significant part of the solution.", "389363": "AmirH is spot on. The special prize’s award will be at the host’s discretion based on the principle of whether you are demonstrating you can win this competition using the best of R, a rich language which offers extensive possibility and inventiveness. So while you may employ some interoperability capabilities of the language’s design, you should expect the host to evaluate the award based on whether R’s use is part of the entire model workflow. You should not expect a response with some quantitative measure of exactly how much of a non-R language you may use. The point is whether you have made a submission with this principle and the prize’s intent in mind.", "389520": "AmirH, @Julia Elliott Got it! Thank you both for the responses, I think I understand the situation better now.", "391299": "julia \n\n1. Is the use of external executables such as libffm as part of the solution is acceptable for the R Prize?\n\n2. If one of the submission is done is R and the other is not, would the R submission alone qualify for the R Prize?\n\nThanks for the clarification!" }, "source": "meta" }