Does the algorithm decode sarcasm?

Does the algorithm decode sarcasm?

The short answer is: it's not necessary for the purpose of doing anthropology, because we're not concerned with verbatims but rather the topics within them.

Our algorithm extracts topics (in the broader context of a search) and quantifies and surfaces the most dominant topics that make up the consumers mental model in the context of a search.

Words used for purposes of humor and sarcasm never make their way into the dominant "solar system" owing to the fact that they do not occur as often or with the same consistency and therefore do not dominate a culture.

They always get weeded out and remain outside the boundaries, further away from the sun. E.g. a search for "sustainable products" is made up of more than 30K topics within it. We focus on the top 5000 (at the most) to make sense of it and examine where the culture is headed.

The noise usually sits in the bottom 5% of the culture, which is where elements of sarcasm and meme-like behavior often play a role.