Results for "drift"
Concept Drift
IntermediateThe relationship between inputs and outputs changes over time, requiring monitoring and model updates.
This term describes a situation where the relationship between the input data and the output predictions changes over time. For example, a model predicting stock prices might become less accurate if market conditions shift significantly. It's like a student who learns a subject but finds that the...
Concept Drift
Intermediate
The relationship between inputs and outputs changes over time, requiring monitoring and model updates.
Prediction Drift
Intermediate
Shift in model outputs.
Data Drift
Intermediate
Shift in feature distribution over time.
Monitoring
Intermediate
Observing model inputs/outputs, latency, cost, and quality over time to catch regressions and drift.