waffles_ts

A command-line tool for time series analysis. Here's the usage information:

Full Usage Information
[Square brackets] are used to indicate required arguments.
<Angled brackets> are used to indicate optional arguments.

waffles_ts [command]
   Analyze and extrapolate time-series data.
   train [dataset] <options>
      Trains an instance of Neural Decomposition on a time-series dataset.
      [dataset]
         The filename of the dataset.
      <options>
         -regularization [value]
            Specify the L1 regularization amount.
         -learningRate [value]
            Specify the learning rate.
         -linearUnits [value]
            Specify the number of linear units to use to augment sinusoid
            units.
         -softplusUnits [value]
            Specify the number of softplus units to use to augment sinusoid
            units.
         -sigmoidUnits [value]
            Specify the number of sigmoid units to use to augment sinusoid
            units.
         -epochs [value]
            Specify the number of epochs to train the model.
         -filterLogarithm
            Use a logarithmic pre- and post-processing step.
         -features [dataset]
            Specify a features matrix instead of generating one automatically.
            Must be a single-column matrix in arff format.
   extrapolate [model-file] <options>
      Use a trained model to extrapolate time-series data.
      <options>
         -start [value]
            Specify where to begin extrapolation. 0 represents the beginning of
            the training data, 1 represents the beginning of the testing data.
         -length [value]
            Specify how far into the future to extrapolate. 1 represents the
            length of the training data.
         -step [value]
            Specify how tight to make predictions.
         -outputFeatures
            Include a features column in the output matrix.
         -features [dataset]
            Specify a features matrix instead of generating one automatically.
            Must be a single-column matrix in arff format. If this is set, then
            the start, length, step, and outputFeatures flags will be ignored.