Back to all projects

Cannabis Recommendation System

A hybrid recommendation system combining AI with multi-vector similarity matching for cannabis applications

Key Features

  • Dramatically reduced API token usage (90% reduction)
  • Increased scalability with local strain database for matching
  • Improved accuracy through mathematical precision with ML algorithms
  • Better performance through sparse matrix optimizations and caching
  • Multi-vector similarity calculations across attribute categories
  • Diversity algorithms to prevent 'filter bubbles' in recommendations
  • Negative correlation insights to avoid contradictory effects
View Code