pangeo_fish.hmm.estimator.CachedEstimator.decode

CachedEstimator.decode(X, states=None, *, mode='viterbi', spatial_dims=None, temporal_dims=None, progress=False, additional_quantities=['distance', 'speed'])

Decode the state sequence from the selected model and the data

Parameters:

• X (xarray.Dataset) – The emission probability maps. The dataset should contain these variables:

  • pdf, the emission probabilities

  • mask, a mask to select ocean pixels

  • initial, the initial probability map

  • final, the final probability map (optional)

• states (xarray.Dataset, optional) – The precomputed state probability maps. The dataset should contain these variables:

  • states, the state probabilities

• mode (str or list of str, default: "viterbi") – The decoding method. Can be one of:

  • "mean": use the centroid of the state probabilities as decoded state

  • "mode": use the maximum of the state probabilities as decoded state

  • "viterbi": use the viterbi algorithm to determine the most probable states

  If a list of methods is given, decode using all methods in sequence.

• additional_quantities (None or list of str, default: [``”distance”, ``"speed"]) – Additional quantities to compute from the decoded tracks. Use None or an empty list to not compute anything.

  Possible values are:

  • "distance": distance to the previous track point in [km]

  • "speed": average speed for the movement from the previous to the current

    track point, in [km/h]

• spatial_dims (list of hashable, optional) – The spatial dimensions of the dataset.

• temporal_dims (list of hashable, optional) – The temporal dimensions of the dataset.