Unverified50% confidenceDecision RuleExact time
Distance- or gradient-based algorithms (KNN, SVM, logistic regression, neural networks) are sensitive to feature scale and require feature scaling.
1
Sources
50%
Confidence
Long-term
Relevance
9/12/2026
First Seen
Sources
Related Entities
Related Claims
Unverified神经网络重要性度量的成熟技术路线包括梯度×输入、积分梯度、SHAP 值、LayerCAM 等73% similarUnverifiedSVM finds the optimal hyperplane by maximizing the classification margin, and the kernel trick can implicitly map data to a high-dimensional space to handle non-linearly separable problems.69% similarUnverifiedSAOT采用的人体姿态估计算法脱胎于计算机视觉深度学习研究，核心是通过卷积神经网络（CNN）提取特征并经三维重建算法合成球员空间坐标69% similarUnverified解决细粒度视觉识别近邻歧义的主流技术路径包括注意力机制、部件检测网络和度量学习68% similarUnverified现代行人重识别方法以深度度量学习为主流，通过三元组损失或对比损失训练孪生网络，将同一目标的不同图像映射至特征空间中相近位置66% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/910735API
curl https://kongchang.com/api/v1/knowledge/claims/910735MCP
get_claim(id=910735)