absolang

Absolang

A sentiment analysis library for detecting absolutist language.

Quickstart

Installation::

$ pip install absolang
$ python -m spacy download en_core_web_sm

Determining absolutist index for text::

>>> from absolang import absolutist, absolutist_index
>>> absolutist_index("The bigger dog is running.")
0.0
>>> absolutist("The bigger dog is running.")
False
>>> absolutist_index("He was completely bowled over.")
0.2
>>> absolutist("He was completely bowled over.")
True

Algorithm

• Parse text into tokens using Spacy's en_core_web_sm language model.
• Count the number of word tokens (a token is considered a word if it consists solely of characters from the alphabet).
• Count the number of absolutist word tokens (a token is considered absolutist if its stem word is in the dictionary of absolutist words and it is not preceded by a negation, modifier or interjection).
• The absolutist index is the number of absolutist words divided by the total number of words.
• Text is considered absolutist if the index is greater than 1.1 percent.

Caveats

• The frequency of absolutist words in control texts (ones written by people presumed not to suffer from anxiety or depression more than the average person) is about 1%, so one needs a few hundred words of texts before results start becoming meaningful.

References

• In an Absolute State: Elevated Use of Absolutist Words Is a Marker Specific to Anxiety, Depression, and Suicidal Ideation <http://journals.sagepub.com/doi/pdf/10.1177/2167702617747074>_

• FCA Occasional Paper No. 8: Consumer Vulnerability <https://www.fca.org.uk/publications/occasional-papers/occasional-paper-no-8-consumer-vulnerability>_