Contextual Bandit (LinUCB)
OptimizePrice: $0.02/callLatency: <1msComplexity: O(d^2 * k)
Quick scenarios
Make context-aware decisions using feature vectors. The bandit learns which arm performs best given the current context (time of day, energy level, etc.).
Click "Run Algorithm" to see results
cURL
curl -X POST https://oraclaw-api.onrender.com/api/v1/optimize/contextual-bandit \
-H "Content-Type: application/json" \
-d '{"arms":[{"id":"email","name":"Email campaign"},{"id":"social","name":"Social media"},{"id":"search","name":"Search ads"}],"context":[0.8,0.6,0.3],"history":[{"armId":"email","reward":0.7,"context":[0.9,0.5,0.2]},{"armId":"social","reward":0.4,"context":[0.3,0.8,0.7]}],"alpha":1}'Input Schema
arms: Array of {id, name}
context: number[] (feature vector)
history: Array of {armId, reward, context}
alpha: number (exploration parameter)
Output Fields
selectedscoreexpectedRewardconfidenceWidthalgorithm