Unverified50% confidenceFactExact time
Data drift refers to a change in the distribution of input features P(X) while the relationship between features and labels P(Y|X) remains unchanged; concept drift is when P(Y|X) itself changes.
1
Sources
50%
Confidence
Long-term
Relevance
10/25/2026
First Seen
Sources
Related Entities
Related Claims
VerifiedConcept drift refers to the mapping relationship between input data and target labels changing over time, which is different from data drift where only the input distribution changes.79% similarUnverified概念漂移是数据底层分布随时间发生的自然变化，而数据投毒是攻击者蓄意制造的分布偏移，两者在实际系统中往往同时存在且难以区分67% similarUnverifiedDistribution shift includes three common types: covariate shift (the input distribution changes but the conditional distribution does not), label shift (the label marginal distribution changes) and concept drift (the conditional distribution changes over time).66% similarVerified数据漂移（Data Drift）指生产环境中输入数据的统计分布随时间发生变化，导致模型接收数据与训练时分布不匹配66% similarUnverifiedData drift is a correlational signal rather than a causal signal: distribution shifts do not necessarily cause model errors, and model errors are not necessarily accompanied by detectable drift65% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1096446API
curl https://kongchang.com/api/v1/knowledge/claims/1096446MCP
get_claim(id=1096446)