Reference
Glossary
Definitions of terms used in TheKnow. Learn the language of personalized restaurant recommendations.
- 01Taste match
- A TheKnow user whose restaurant ratings correlate with yours. When they love a restaurant, you're likely to love it too. Taste matches are identified using Pearson correlation, and the strength of the match is displayed as a percentage. Finding taste matches is the core of how TheKnow delivers personalized recommendations.
- See also:Taste twinPearson correlation
- 02Taste twin
- A taste match with exceptionally strong alignment—typically 70% correlation or higher. Taste twins have nearly identical restaurant preferences, making their recommendations highly reliable. When a taste twin discovers a great restaurant, you can trust their opinion almost as much as your own.
- See also:Taste match
- 03TES score
- Taste Experience Score. A 0-999 metric that measures your engagement and expertise on TheKnow. Your TES score reflects how many restaurants you've rated, how many cuisines you've explored, and how many taste matches you've connected with. A higher TES indicates a more developed taste profile and more reliable recommendations.
- See also:Profile tierTaste graph
- 04Taste graph
- A visual representation of your dining personality across 6 dimensions: cuisine variety, price range, atmosphere preferences, dining frequency, adventurousness, and quality focus. Your taste graph helps you understand your own patterns and shows others what kind of diner you are. It's generated from your ratings and preferences.
- See also:TES score
- 05For You score
- A predicted rating for a restaurant you haven't visited yet. TheKnow calculates this by analyzing how your taste matches rated the restaurant, weighted by match strength. A For You score of 4.2 means your taste matches collectively suggest you'll enjoy it at about 4.2 out of 5 stars.
- See also:Taste matchPersonalized recommendations
- 06Pearson correlation
- The statistical method TheKnow uses to measure taste alignment between users. Pearson correlation analyzes whether two users' ratings move together: when one rates highly, does the other? A positive correlation means similar taste; the higher the number, the stronger the match. This is the same technique used in recommendation systems for music, movies, and products.
- See also:Taste matchAlgorithm
- 07Profile tier
- A trading-card style designation that reflects your engagement level on TheKnow. Tiers progress from Standard (just starting) to Silver (regular rater) to Gold (frequent contributor) to Holographic (top-tier taste expert). Your tier is determined by your TES score and overall activity.
- See also:TES score
- 09Personalized recommendations
- Restaurant suggestions tailored specifically to your taste profile. Unlike crowd-based apps like Yelp that show the same recommendations to everyone, TheKnow's recommendations are unique to you. They're based on what your taste matches love, weighted by how strong your match is with each person.
- See also:For You scoreTaste match
- 10Crowd ratings
- Average scores from a large, anonymous group of reviewers (like on Yelp or Google). TheKnow was built as an alternative to crowd ratings because a restaurant's average score doesn't account for individual taste. A place with 3.5 stars from the crowd might be a 5 for you if your taste matches love it.
- See also:Personalized recommendations
- 11Rating scale
- TheKnow uses a 0 to 5 star scale with 0.5 increments, where 5 represents exceptional and 0 represents terrible. Honest ratings—including low ones—help the algorithm work better. Don't be afraid to rate something poorly if you didn't enjoy it.
- See also:Shared ratings
- 12Time decay
- A factor in TheKnow's algorithm that gives more weight to recent ratings. A restaurant you rated 5 stars last month is more relevant than one you rated 5 stars three years ago. Time decay ensures recommendations stay current as both your taste and restaurants evolve.
- See also:AlgorithmPearson correlation
- 13Confidence weighting
- A method TheKnow uses to adjust predictions based on how much data supports them. A For You score based on 10 taste matches is more confident than one based on 2. More shared ratings between you and a match increases confidence in that match's correlation.
- See also:For You scoreShared ratings