Spatio-Temporal Beam-Level Traffic Forecasting Challenge by ITU
12 000 CHF
Completed (over 1 year ago)
Forecast
732 joined
169 active
Start
Jul 24, 24
Close
Oct 11, 24
Reveal
Oct 11, 24
Still scoring after 18 hours. Why do we have so many issues with submissions ?
Help · 25 Jul 2024, 15:52 · 6
Please for those who submitted successly. How much time does it take to score ?
Any answer will be appreciated.
Update (26/07/2024) : I don't know the SampleSubmission.csv works with submissions for the 4 for files. But When I try to submit only for traffic_DLThpVol, the file has not been scored since. And when I try the 4 files, I get a score.
Update (27/07/06) But I tried a sub today with the 4 files (27/07/2024) and IT IS STILL SCORING for 2 hours now. We were told that submission will take ~50 mins.
Hello, it seems that you made a valid submission yourself. Could you kindly share your findings with respect to the question you asked.
My submissions took nearly 30 mins. I think there are many misleading informations in the Info section. My submission which is still scoring only provided submission for traffic_DLThpVol only. But when I submit prediction for the 4 files, I get a score. SO IN THIS CHALLENGE WE MUST PREDICT FOR ALL FILES.
Yep, the info page is not clear, it suggests we have to submit traffic (DLThp) only in wide format, but looking at the submission file we need to submit all 4 in long format.
Just for reference, I also have a submission now running for over two hours, same number of rows as sample.
@Amy_Bray, @Zindi Can you please clarify this, before we start deep diving? I'm just making sure that there are no changes in the future.
According to Data (About Section):
"We remind the participants that the objective is to forecast future values of traffic volume (DLThpVol)."
If this was true, then we needed to submit 2880*24*7*2 rows in the submission file which is equal to 967680. Here 2880 are number of traffic beams, 24 is hours in a day, 7 is days in a week and 2 are number of test weeks.
But According to Sample submission:
We need to submit 3870720 rows, which is equal to 2880*24*7*2*4
The last 4 means all the 4 different metrics require forecast instead of just DLThpVol which is not what is said in info section.
And if this is the case it becomes a problem of forecasting 2880*4 series instead of just 2880 series, which has some impact in the solution design, so an earlier response would be appreciated.
Thanks,
Krishna
They don't even respond. I think that challenge participants need more consideration from @Zindi. As I said in another discussion, I'm considering leaving Zindi and give up on ongoing challenges.
No don't leave. It is fun to compete when you have tough competition from others :)
As for this one I have decided to start only if and once we get a clarification, otherwise the experiments and submissions go to a waste.