Python RecurrenceNetwork.embed_time_series Examples

Example #1

File: recurrence_analysis_speleo_raw.py Project: yangxhcaf/pyregimeshifts

    #
    #  Calculate significance levels for network measures
    #

    print "Calculating significance levels based on", N_ENSEMBLE, "surrogates..."

    #  Initialize progress bar
    progress = progressbar.ProgressBar().start()

    #  Create a copy of data for generating surrogates from
    surrogate_data = data.copy()

    if SHUFFLE_EMBEDDED:
        #  Get embedding of full time series
        surrogate_embedding = rec_net.embed_time_series(
            surrogate_data, DIM, TAU)

    #  Prepare stuff
    local_surrogate_result = {}

    for measure in symbols.keys():
        local_surrogate_result[measure] = np.empty(N_ENSEMBLE)

    for j in xrange(N_ENSEMBLE):
        if SHUFFLE_EMBEDDED:
            #  Shuffle embedded time series along time axis, that is, whole
            #  embedded state vectors are shuffled around.
            permuted_indices = np.random.permutation(
                surrogate_embedding.shape[0])

            #  Use the first T state vectors from the shuffled and embedded