From Anecdote to Algorithm: How Data Science and Machine Learning are Redefining Cannabis Classification and Reshaping the Consumer Experience

The recreational cannabis market currently relies on unreliable phenotypic classification methods (e.g., strain names and lineage, “potency,” and morphology) that fail to correlate accurately with the plant’s true chemical complexity or the resulting consumer experience. In its simplest categorization, cannabis is relegated to indica, sativa, or hybrid, allegedly providing predictable effects across all consumers However, there is no standardization to the various strain names or indica, sativa, and hybrid designations. This long-standing classification issue, perpetuated by historical research prohibition and cultural precedent, has led to inaccurate and often ineffective consumer recommendations, as well as inefficient market operations.
Terpenes are well known to be a class of compounds that influence the effects of a particular cannabis plant. Organoleptic techniques have been employed informally by cannabis consumers and professionals, similar to the food and beverage industry, as a means to ensure quality, facilitate education, and enhance consumer experience. Combining organoleptic knowledge with mining years of accumulated analytical cannabis data (including comprehensive cannabinoid and terpene profiles) and employing machine learning algorithms, we successfully identified chemotype-based clusters that provide a more robust and predictive classification system versus traditional nomenclature. The resulting system identified six major categories of cannabis plants and use colors to correspond to each category. Because the categories are based on analytical data, this categorization is not only improving the reliability of consumer consultations, but is also influencing professional retail buying strategies and providing cultivators and processors with actionable data on terroir. This work establishes a critical paradigm shift away from anecdotal classification, offering a scientific framework for cannabis discernment.