Modelling NPTS model

[parimuns]

I am trying to model NPTS using Gluonts. Please guide me on how to get the probabilistic forecasting results using NPTSEstimator.

'from gluonts.model.npts import NPTSEstimator
from gluonts.trainer import Trainer
from gluonts.model.npts import NPTSPredictor
Predictor1=NPTSPredictor(freq='H', prediction_length=24, context_length= 48, kernel_type='exponential',use_seasonal_model= True, use_default_time_features = True)
predictor=Predictor1.predict(dataset=train_ds,num_samples=100)
test_ds = ListDataset([{FieldName.TARGET: target, 
                        FieldName.START: start,
                        FieldName.FEAT_DYNAMIC_REAL: [fdr],
                       } 
                       for (target, start,fdr) in zip(target, 
                                                            start, 
                                                            feat_dynamic_real, 
                                                            )],
                     freq=freq)
from gluonts.evaluation.backtest import make_evaluation_predictions
forecast_it, ts_it = make_evaluation_predictions(
    dataset=test_ds,  # test dataset
    predictor=predictor,  # predictor
    num_samples=100,  # number of sample paths we want for evaluation
)

I get error

AttributeError                       Traceback (most recent call last)
<ipython-input-25-c98cb60cf47a> in <module>
      3     dataset=test_ds,  # test dataset
      4     predictor=predictor,  # predictor
----> 5     num_samples=100,  # number of sample paths we want for evaluation
      6 )

c:\users\appdata\local\programs\python\python37\lib\site-packages\gluonts\evaluation\backtest.py in make_evaluation_predictions(dataset, predictor, num_samples)
     60     """
     61 
---> 62     prediction_length = predictor.prediction_length
     63     freq = predictor.freq
     64 

AttributeError: 'generator' object has no attribute 'prediction_length'

[jaheba]

You set predictor to Predictor1.predict(...). predict() returns an iterator over the predictions.

I think what you want is to pass Predictor1 to make_evaluation_predictions instead.

[parimuns]

Got it.Thanks