Hi,
when analyzing the data I say that (per room) all the devices produce the same number of people.
when looking only on the 360 samples with the same num-windows, we miss the other 360 samples of other devices with the same timestamp.
Is it possible to get access to the entire data in order to calculate the new features that I require?
My model is looking for some average between the devices in the room.
Sincerely,
Lev.
Hi,
The prediction should be based on a single independent window. We recommend examining inter-room and intra-room correlation as part of the train data analysis, however, please note that the public and private test sets do not include the Room_Num feature.
MAFAT Challenge Team