Unverified50% confidenceFactExact time
Scaffold Split groups molecules according to their Bemis-Murcko scaffolds, ensuring that all molecules sharing the same scaffold appear only on the training set or test set side, in order to simulate the ability to predict unfamiliar chemical scaffolds.
1
Sources
50%
Confidence
Long-term
Relevance
9/12/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedList-based processing capability enables protein structure prediction to be incorporated into automated R&D pipelines, rather than being executed as one-off computational experiments.62% similarUnverified对Hadamard乘积进行聚类分析可以提取神经元检测的全部语义模式，产生清晰的单义性簇57% similarUnverifiedMolCLR constructs positive and negative sample pairs by applying different data augmentations to molecular graphs, while Grover and GraphMAE adopt masked pre-training strategies to learn molecular representations.57% similarUnverifiedThe SMOTE algorithm achieves oversampling by interpolating between minority class samples to generate synthetic samples.56% similarUnverifiedWhen processing molecular tasks, Scaffold Split should be used instead of random splitting; otherwise, the model's performance will be inflated due to data leakage.56% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/914907API
curl https://kongchang.com/api/v1/knowledge/claims/914907MCP
get_claim(id=914907)