Easy-Handler for Kaggle Annotated Corpus for Named Entity Recognition

It helps you load the kaggle dataset easily. You first download it from https://www.kaggle.com/abhinavwalia95/entity-annotated-corpus

Usage

load function partitions the dataset to train, develop, test with permutation. The format of each dataset is list of sentence and the format of each sentence is list of tuple. The tuple consists with (word, POS, NER tag).

from kaggle_ner_dataset_handler import load

data_dir = 'path/to/entity-annotated-corpus/'
train, develop, test = load(data_dir)
train[0]  # select first sentence

[('Israel', 'NNP', 'B-geo'),
 ("'s", 'POS', 'O'),
 ('security', 'NN', 'O'),
 ('cabinet', 'NN', 'O'),
 ('has', 'VBZ', 'O'),
 ('approved', 'VBN', 'O'),
 ('a', 'DT', 'O'),
 ...

lovit/kaggle_ner_dataset_handler

Easy-Handler for Kaggle Annotated Corpus for Named Entity Recognition

Usage