-use the p spline to get the smooth data, then the dataset only contains 0 at the start or end because the range of meaured date is slightly different
-the nan in predicted value(from spline) will be change to 0 or the value of the last day
-
data simulation: use SDE model to generate plant growth cureves(4 types limitation and 4 types noise)
data from 'model_comparision/data/simulated_data/no_fixed_Max/' are used for classification