willwade/txt-sms-abbreviations

Dataset Card for TXT/SMS Abbreviations

1539 sms/text speak abbreviations. Note we have parsed through NLTK to give a field of whether its generally recognised a whole word. NB: Numbers are NOT noted as words. Be careful with numbers like 4 2 etc in your training data

Dataset Details

Dataset Description

A curated list of abbreviations commonly used in text messages and gaming chat and other places. We have tried to remove explanations and explicitly put a full expansion of the abbreviation. Be careful - some items are whole words already.

• Curated by: Will Wade
• Language(s) (NLP): English
• License: Apache-2

Uses

Use as a training set for abbreviations or as part of a GEC task

import csv
# Load abbreviations and their expansions from sms.csv into a dictionary
abbreviations_dict = {}
with open('sms.csv', mode='r', encoding='utf-8') as infile:
    reader = csv.reader(infile, delimiter=',')
    next(reader, None)  # Skip the header if present
    for row in reader:
        if len(row) >= 3:
            abbreviation = row[0].strip().upper()
            expansion = row[1].strip()
            is_word = row[2].strip().lower() in ['true', 'yes', '1']  # Convert to boolean
            abbreviations_dict[abbreviation] = (expansion, is_word)
        else:
            print(f"Skipping malformed row: {row}")