labMTsimple: upgrades! hedonometer python programming
As part of an ongoing effort to evaluate different sentiment analysis tools, I've pushed a handful of upgrades to labMTsimple. Namely:
- Python 3 compatible
- New module "speedy," aiming for speed (also using numpy)
- Class-based sentiment object
- ANEW, MPQA, and Liu's dictionaries are included
- Using "marisa-trie" and "datrie" for fast prefix matching
It's all available on github as well as on pypi for installation via pip.
Of course, the vectors produced for the labMT set in english by the speedy module are ordered the same. Also, no other languages for labMT on speedy yet.
Stay tuned for the results of performance tests from these other dictionaries on a variety of corpuses.