Primary competition visual

Spatio-Temporal Beam-Level Traffic Forecasting Challenge by ITU

12 000 CHF
Completed (over 1 year ago)
Forecast
716 joined
171 active
Starti
Jul 24, 24
Closei
Oct 11, 24
Reveali
Oct 11, 24
User avatar
RareGem
Enquiries
Platform · 16 Sep 2024, 21:12 · 1

Why is the submission data length higher than the training data length? Please,how can I handle that?

Discussion 1 answer

Hi RareGem,

Here is my understanding.

The training data “traffic_DLThpVol” contains hourly data for 5 weeks, which amounts to a total of 840 rows (24 hours x 7 days x 5 weeks). The columns represent the data corresponding to each beam, with a total of 2880 columns (30 base stations x 3 cells x 32 beams).

On the other hand, the submission data is the data for the 6th and 11th weeks, arranged vertically. One week of data has 2880 x (24 hours x 7 days) = 483,840 rows, resulting in 967,680 rows for 2 weeks.

17 Sep 2024, 03:36
Upvotes 2