Hikematch - Recommender System
A hybrid, contextual recommender system for hiking routes
We build a recommender system for hiking routes that considers current contextual data such as weather conditions, time of day, and user location, as well as explicit user preferences. For the semantic recommender component, we designed a system to discover latent topics across the hike descriptions and build a knowledge graph linking hikes based on their topics (topics capture how each hike “feels” to the user).