Back to all projectsView Code
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