`sparknlp_jsl.utils.ner_utils`#

Module Contents#

`evaluate`(true_seqs, pred_seqs[, verbose])	Evaluate performance of a sequence tagger.
`evaluate_conll_file`(fileIterator)	Prints overall performance, as well as performance per chunk type.
`get_charts`(log_file[, threshold])	Plots the figures of metrics ( precision, recall, f1) vs epochs.
`loss_plot`(log_path)	Plots the figure of loss vs epochs.

evaluate(true_seqs: List[str], pred_seqs: List[str], verbose=True)#

Evaluate performance of a sequence tagger.

If verbose, returns overall performance, as well as performance per chunk type; otherwise, simply returns overall precision, recall, f1 scores

Parameters:

true_seqs (List[str]) – a list of true tags
pred_seqs (List[str]) – a list of predicted tags
verbose (bool) – whether to print overall performance, as well as performance per chunk type (default is True).

evaluate_conll_file(fileIterator)#: Prints overall performance, as well as performance per chunk type.

get_charts(log_file: str, threshold: float = 0.0)#

Plots the figures of metrics ( precision, recall, f1) vs epochs.

Parameters:

loss_plot(log_path: str)#

Plots the figure of loss vs epochs.