I know there is a thread already and someone pointed that it is allowed. But i am confused by this two lines in the data section:
"The labels indicate the growth stage of the crop in the image Growth stages indicate the maturity of the wheat plants and are represented as a number from 1 to 7 (mature crop)."
"you must predict the growth stage for the crops in each of these images."
Are we really allowed to use continous prediction? Submission will not generate error whether i submit continuous or not but i dont want to be disqualified from the leaderboard later for this.
yes i have same question and continuous values decrease the mse
@ai...... I don't think you should worry yourself too much on that. The simple trick i will suggest, is to round up these continous "Growth Stages" in your prediction and later submit. Then check if you will still have the same score or not on the Leaderboard when comparing these to your previous submitted continous prediction. That's my candid advise anyway. Good luck....
However, we will all appreciate this if @Zindi can come in on this question that's been begging for answers for a while now in this challenge.
Continuous is allowed - notice the metric. The rationale is that these aren't discrete stages and if a model thinks the wheat is stage 4 or 5 then an output of 4.5 is useful.
Yes, according to the metric it is more reasonable to predict continous values but an offical statement from @Zindi on this issue would be nice.
@Zindi , This is a crucial question to answer. Please do not ruin efforts of the participants, just because question was left unanswered.
Yes, continuous prediction is allowed.