evilometer: Rates arbitrary names based on a pre-rated list of names on some characteristic (ie "evilness")

• Copyright : July-2014, (c) AGPLv3 or later
• Developed : by [email protected] & Nikifors Zacharoff
• License : GNU Affero General Public License v3 or later (AGPLv3+)

Overview

Given a pre-rated list of names on some characteristic, it decomposes them using n_grams and applies information retrieval rating[inv_index]_ to estimate the rating of any other name on that characteristic.

Install:

To install it, assuming you have download the sources, do the usual:

python setup.py install

Or get it directly from the PIP repository:

pip3 install evilometer

Tested with Python 3.4.

Usage:

Fuefit accepts as input 2 vectors of names, the pre-rated "training" set of and the set of names to be rated, like that:

>> import evilometer

>> train_names = {'trendy': 1, 'good':2, 'better':2, 'talon':-5, 'bad_ass':-10}
>> asked_names = {'kolon':1, 'trekking':2, 'trepper':-10}

>> name_scores = evilometer(train_names, asked_names)
>> print_scored_names_sorted(name_scores)

cmdline

>> python -m evilometer prerated.csv asked.txt

Contributors

• Nikifors Zacharoff & Nikias Fontaras: Original idea of the game

Footnotes

[inv_index]

http://nlp.stanford.edu/IR-book/html/htmledition/inverse-document-frequency-1.html