KaiMaffucci/low-resource-word-matcher

A spelling similarity matcher/detector geared toward finding cognates, spelling variants, and misspellings in low-resource and endangered languages. # low-resource-word-matcher A writing similarity matcher/detector for finding cognates and writing variants in low-resource and endangered languages, particularly Cherokee. **Navigation guide** - To view copyright and sovereignty notices, keep scrolling in this README. - To learn how to use the software, scroll even further. - To look at the dataset, testing process, and findings/discussion for the results of the Cherokee version, check the `data` directory (contained within the `data` directory to maintain clarity about copyright). - To look at the code, look at all of the other directories (I recommend starting with `main.py` in the root of this repository). ## Important Copyright Information DATA SOVEREIGNTY AND COPYRIGHT NOTICE: All Cherokee language material in the `data` folder does NOT fall under the MIT license of the rest of the repository. I do NOT own the data, and it is subject to the original copyright holders, detailed below. If any of the respective copyright holders of the data wish for me to take it down or modify it, feel free to email me here, and such matters will be handled immediately. Syllabary writings for all non-clothing words found on the Cherokee Nation of Oklahoma Language Department's posters section of their website here: language.cherokee.org Words on clothing found here: shiyo.org All words were cross-referenced with entries in the Durbin Feeling dictionary here, based on the English translation: cherokeedictionary.net ## Cherokee Sovereignty Notice I am NOT Cherokee NOR indigenous, and I do not claim that identity, nor am I an authority in Cherokee linguistics. Take everything I say with a huge grain of salt. Furthermore, this research project is conducted in accordance with the CARE guidelines for researchers (navigating.art). Importantly, str …