Language Of Drug Addiction



06 Jun 2020 (Last Modified 06 Jun 2020)


Language Of Drug Addiction

Starting with notes from A Computational Analysis of the Language of Drug Addiction. I will expand this into a notebook a la Cosma Shalizi.

Predose evolved into eDrugTrends and iN3.

Drezde has used factorial LDA to group posts on drugs-forum.

A main goal seems to be to infer mental states from patterns of speech and infer drangements of thoses mental states based on other patterns of speech. An Alcoholic Language Corpus exists, but it is only in German.

Issues with interppreting online comments.

  1. Reporting/Selection bias - We expect less experienced users and those with remarkable expericence to be more likely to post.
  2. Contamination - Clandestine drugs have no quality control. The user may believe that he is using one substance, but is, in reality, using another substance or combination of substances.

Infering physiological mental states

Inferring pathological mental states

Latent user attribute prediction.

Many things on this page still need links. How to get corpus from Svarappa?

  1. PREDOSE

To Read

  1. Wordnet-affect