Sorry for this very simple question but I couldn't find the answer for sure.
\nI noticed there are timestamps in provided sample_submission.csv that are not in test set.
\nAnd there are timestamps in test set that are not in sample_submission.csv
So my question is : which site / path / timestamp must be submitted, for final results ?
\nShould we only submit those in provided sample_submission.csv , which is 10133 lines ?
Or should we also submit site / path / timestamp that are in test set ?
\nI've seen that public notebooks only submit the 10133 lines in sample_submission.csv, but I wanted to be sure that this is the same requirement for final submission.
\nIn the evaluation instructions, it's written : \"For each site_path_timestamp row in the test set, you must predict […]\" which is not exactly what seems to be done there.
\nThanks for your help
", "messages": [ { "id": "1282352", "postDate": "04/23/2021 20:37:16", "content": "Sorry for this very simple question but I couldn't find the answer for sure.
\nI noticed there are timestamps in provided sample_submission.csv that are not in test set.
\nAnd there are timestamps in test set that are not in sample_submission.csv
So my question is : which site / path / timestamp must be submitted, for final results ?
\nShould we only submit those in provided sample_submission.csv , which is 10133 lines ?
Or should we also submit site / path / timestamp that are in test set ?
\nI've seen that public notebooks only submit the 10133 lines in sample_submission.csv, but I wanted to be sure that this is the same requirement for final submission.
\nIn the evaluation instructions, it's written : \"For each site_path_timestamp row in the test set, you must predict […]\" which is not exactly what seems to be done there.
\nThanks for your help
", "rawMarkdown": "Sorry for this very simple question but I couldn't find the answer for sure.\n\nI noticed there are timestamps in provided sample_submission.csv that are not in test set.\nAnd there are timestamps in test set that are not in sample_submission.csv\n\nSo my question is : which site / path / timestamp must be submitted, for final results ?\nShould we only submit those in provided sample_submission.csv , which is 10133 lines ?\n\nOr should we also submit site / path / timestamp that are in test set ?\n\nI've seen that public notebooks only submit the 10133 lines in sample_submission.csv, but I wanted to be sure that this is the same requirement for final submission.\n\nIn the evaluation instructions, it's written : \"For each site_path_timestamp row **in the test set**, you must predict [...]\" which is not exactly what seems to be done there.\n\nThanks for your help", "votes": null }, { "id": "1282368", "postDate": "04/23/2021 21:09:31", "content": "Yes, only submit the 10133 rows in the sample_submission.csv file. So your csv that you submit should have all the same rows as sample_submission.csv, but you change the floor, x, and y values for each line
", "rawMarkdown": "Yes, only submit the 10133 rows in the sample_submission.csv file. So your csv that you submit should have all the same rows as sample_submission.csv, but you change the floor, x, and y values for each line", "votes": null }, { "id": "1282641", "postDate": "04/24/2021 06:28:00", "content": "Thanks for jour answer
\nI was afraid that the 10133 was just a subset for evaluating public scores before the competition ends,
\nThe evaluation instructions saying to use each row in the test set made it confusing
Ah, ok. Yes, the 10133 files are all of the test waypoints, and the public leader board is some 15% of those.
", "rawMarkdown": "Ah, ok. Yes, the 10133 files are all of the test waypoints, and the public leader board is some 15% of those.", "votes": null }, { "id": "1290249", "postDate": "05/01/2021 18:38:58", "content": "Hey Guys
\nI am a Data Science Masters student and I am doing this competition as a part of my ML assignment. My assignment submission deadline is the day after tomorrow and I can't even make a single submission. I analysed and found that not all timestamps in sample_submission.csv are present in the test data. How to deal with this?
\nCan anybody help me please?
Thanks in advance
", "rawMarkdown": "Hey Guys\n\nI am a Data Science Masters student and I am doing this competition as a part of my ML assignment. My assignment submission deadline is the day after tomorrow and I can't even make a single submission. I analysed and found that not all timestamps in sample_submission.csv are present in the test data. How to deal with this? \nCan anybody help me please? \n\nThanks in advance", "votes": null }, { "id": "1290256", "postDate": "05/01/2021 18:46:43", "content": "You're correct - the test data files don't include any waypoints. Instead, you'll want to read the site_uuid, path_uuid and timestamp entries from the sample submission .csv, and use those site/path/timestamps as your waypoints
\nIt's kind of confusing since the waypoints are in the training data and not in the test data txt files - but just pull those rows from the sample submission file and use those site/path/timestamps for your own submission.
\nHope that helps!
", "rawMarkdown": "You're correct - the test data files don't include any waypoints. Instead, you'll want to read the site_uuid, path_uuid and timestamp entries from the sample submission .csv, and use those site/path/timestamps as your waypoints\n\nIt's kind of confusing since the waypoints are in the training data and not in the test data txt files - but just pull those rows from the sample submission file and use those site/path/timestamps for your own submission.\n\nHope that helps!", "votes": null }, { "id": "1290270", "postDate": "05/01/2021 19:00:02", "content": "Hey @chris62, Thanks for the prompt reply :-), Again the question is that some of the timestamp (exactly 479) which are present in sample_submission.csv, are not present in any of the test files. So how would I get predictions for those cases which has these timestamps?
", "rawMarkdown": "Hey @chris62, Thanks for the prompt reply :-), Again the question is that some of the timestamp (exactly 479) which are present in sample_submission.csv, are not present in any of the test files. So how would I get predictions for those cases which has these timestamps?", "votes": null }, { "id": "1290276", "postDate": "05/01/2021 19:03:35", "content": "Ah - don't worry about there being data (sensor or wifi data) at exactly the same time - that doesn't matter.
\nTake a look at one of the training files - see the \"TYPE_WAYPOINT\" lines? those are the lines missing from the test files (at the relative timestamps from the sample submission csv file)
\nSo you could actually add some TYPE_WAYPOINT lines into the test files if you'd like - but don't have to - your code just has to know that those timestamps are the ones you're predicting for
", "rawMarkdown": "Ah - don't worry about there being data (sensor or wifi data) at exactly the same time - that doesn't matter.\n\nTake a look at one of the training files - see the \"TYPE_WAYPOINT\" lines? those are the lines missing from the test files (at the relative timestamps from the sample submission csv file)\n\nSo you could actually add some TYPE_WAYPOINT lines into the test files if you'd like - but don't have to - your code just has to know that those timestamps are the ones you're predicting for", "votes": null }, { "id": "1290296", "postDate": "05/01/2021 19:33:34", "content": "One of the difficulties in this competition is that you don't have input variables (nor target) neatly provided in front of each timestamps… You have to look for closest values in time.
\nFor missing timestamps in submission set, I'm looking for closest lines in time in test set.
\nFor example if 2 values are separated from 50ms each, that's nearly equivalent, you can use them together.
\nThis has been a huge difficulty for me in this competition, since I've wanted to to all data preparation myself and not rely on other's wifi features etc… But that took me way too much time to do
", "rawMarkdown": "One of the difficulties in this competition is that you don't have input variables (nor target) neatly provided in front of each timestamps... You have to look for closest values in time.\n\nFor missing timestamps in submission set, I'm looking for closest lines in time in test set.\n\nFor example if 2 values are separated from 50ms each, that's nearly equivalent, you can use them together.\n\nThis has been a huge difficulty for me in this competition, since I've wanted to to all data preparation myself and not rely on other's wifi features etc... But that took me way too much time to do", "votes": null } ], "comments": [ { "id": 1282368, "author_name": "chris62", "author_url": "", "post_date": "04/23/2021 21:09:31", "content": "Yes, only submit the 10133 rows in the sample_submission.csv file. So your csv that you submit should have all the same rows as sample_submission.csv, but you change the floor, x, and y values for each line
", "votes": null, "replies": [] }, { "id": 1282641, "author_name": "franoisboyer", "author_url": "", "post_date": "04/24/2021 06:28:00", "content": "Thanks for jour answer
\nI was afraid that the 10133 was just a subset for evaluating public scores before the competition ends,
\nThe evaluation instructions saying to use each row in the test set made it confusing
Ah, ok. Yes, the 10133 files are all of the test waypoints, and the public leader board is some 15% of those.
", "votes": null, "replies": [] } ] }, { "id": 1290249, "author_name": "amitkumargangwar", "author_url": "", "post_date": "05/01/2021 18:38:58", "content": "Hey Guys
\nI am a Data Science Masters student and I am doing this competition as a part of my ML assignment. My assignment submission deadline is the day after tomorrow and I can't even make a single submission. I analysed and found that not all timestamps in sample_submission.csv are present in the test data. How to deal with this?
\nCan anybody help me please?
Thanks in advance
", "votes": null, "replies": [ { "id": 1290256, "author_name": "chris62", "author_url": "", "post_date": "05/01/2021 18:46:43", "content": "You're correct - the test data files don't include any waypoints. Instead, you'll want to read the site_uuid, path_uuid and timestamp entries from the sample submission .csv, and use those site/path/timestamps as your waypoints
\nIt's kind of confusing since the waypoints are in the training data and not in the test data txt files - but just pull those rows from the sample submission file and use those site/path/timestamps for your own submission.
\nHope that helps!
", "votes": null, "replies": [] } ] }, { "id": 1290270, "author_name": "amitkumargangwar", "author_url": "", "post_date": "05/01/2021 19:00:02", "content": "Hey @chris62, Thanks for the prompt reply :-), Again the question is that some of the timestamp (exactly 479) which are present in sample_submission.csv, are not present in any of the test files. So how would I get predictions for those cases which has these timestamps?
", "votes": null, "replies": [ { "id": 1290276, "author_name": "chris62", "author_url": "", "post_date": "05/01/2021 19:03:35", "content": "Ah - don't worry about there being data (sensor or wifi data) at exactly the same time - that doesn't matter.
\nTake a look at one of the training files - see the \"TYPE_WAYPOINT\" lines? those are the lines missing from the test files (at the relative timestamps from the sample submission csv file)
\nSo you could actually add some TYPE_WAYPOINT lines into the test files if you'd like - but don't have to - your code just has to know that those timestamps are the ones you're predicting for
", "votes": null, "replies": [] }, { "id": 1290296, "author_name": "franoisboyer", "author_url": "", "post_date": "05/01/2021 19:33:34", "content": "One of the difficulties in this competition is that you don't have input variables (nor target) neatly provided in front of each timestamps… You have to look for closest values in time.
\nFor missing timestamps in submission set, I'm looking for closest lines in time in test set.
\nFor example if 2 values are separated from 50ms each, that's nearly equivalent, you can use them together.
\nThis has been a huge difficulty for me in this competition, since I've wanted to to all data preparation myself and not rely on other's wifi features etc… But that took me way too much time to do
", "votes": null, "replies": [] } ] } ], "raw_markdown_by_id": { "1282352": "Sorry for this very simple question but I couldn't find the answer for sure.\n\nI noticed there are timestamps in provided sample_submission.csv that are not in test set.\nAnd there are timestamps in test set that are not in sample_submission.csv\n\nSo my question is : which site / path / timestamp must be submitted, for final results ?\nShould we only submit those in provided sample_submission.csv , which is 10133 lines ?\n\nOr should we also submit site / path / timestamp that are in test set ?\n\nI've seen that public notebooks only submit the 10133 lines in sample_submission.csv, but I wanted to be sure that this is the same requirement for final submission.\n\nIn the evaluation instructions, it's written : \"For each site_path_timestamp row **in the test set**, you must predict [...]\" which is not exactly what seems to be done there.\n\nThanks for your help", "1282368": "Yes, only submit the 10133 rows in the sample_submission.csv file. So your csv that you submit should have all the same rows as sample_submission.csv, but you change the floor, x, and y values for each line", "1282641": "Thanks for jour answer\nI was afraid that the 10133 was just a subset for evaluating public scores before the competition ends,\nThe evaluation instructions saying to use each row in the test set made it confusing", "1282935": "Ah, ok. Yes, the 10133 files are all of the test waypoints, and the public leader board is some 15% of those.", "1290249": "Hey Guys\n\nI am a Data Science Masters student and I am doing this competition as a part of my ML assignment. My assignment submission deadline is the day after tomorrow and I can't even make a single submission. I analysed and found that not all timestamps in sample_submission.csv are present in the test data. How to deal with this? \nCan anybody help me please? \n\nThanks in advance", "1290256": "You're correct - the test data files don't include any waypoints. Instead, you'll want to read the site_uuid, path_uuid and timestamp entries from the sample submission .csv, and use those site/path/timestamps as your waypoints\n\nIt's kind of confusing since the waypoints are in the training data and not in the test data txt files - but just pull those rows from the sample submission file and use those site/path/timestamps for your own submission.\n\nHope that helps!", "1290270": "Hey @chris62, Thanks for the prompt reply :-), Again the question is that some of the timestamp (exactly 479) which are present in sample_submission.csv, are not present in any of the test files. So how would I get predictions for those cases which has these timestamps?", "1290276": "Ah - don't worry about there being data (sensor or wifi data) at exactly the same time - that doesn't matter.\n\nTake a look at one of the training files - see the \"TYPE_WAYPOINT\" lines? those are the lines missing from the test files (at the relative timestamps from the sample submission csv file)\n\nSo you could actually add some TYPE_WAYPOINT lines into the test files if you'd like - but don't have to - your code just has to know that those timestamps are the ones you're predicting for", "1290296": "One of the difficulties in this competition is that you don't have input variables (nor target) neatly provided in front of each timestamps... You have to look for closest values in time.\n\nFor missing timestamps in submission set, I'm looking for closest lines in time in test set.\n\nFor example if 2 values are separated from 50ms each, that's nearly equivalent, you can use them together.\n\nThis has been a huge difficulty for me in this competition, since I've wanted to to all data preparation myself and not rely on other's wifi features etc... But that took me way too much time to do" }, "source": "meta" }