[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"$fRNmUAlELd5Apw0Qm10TCalH5M1mDxMSj4FthilhmFCI":3},{"answer":4,"createTime":5,"id":6,"options":7,"origin":13,"question":20,"related":21,"source":31,"type":97},[],"2026-04-08 10:51:35",336173408,[8,9,10,11,12],"损失函数的形状","优化器的类型","训练数据的规模","模型的层数","激活函数的选择",{"count":14,"courseId":15,"courseImg":16,"courseName":17,"workId":18,"workName":19},10,"53e1d2ef4961cca8eea3e23969ad2cb9","https:\u002F\u002Ftihai-oss-cloud.itihey.com\u002Fimg\u002F03a579384a6dc297c89809b582fcc767.png","默认课程","c688bbc5da5a4d1492e7ab7f9953487e","第二章习题","以下哪些因素可能会影响模型的学习率选择方法",[22,33,42,51,60,69,78,87,98,108],{"answer":23,"createTime":5,"id":24,"options":25,"question":30,"source":31,"type":32},[],336173399,[26,27,28,29],"NFL定理指出所有机器学习算法在所有问题上的性能是相同的","NFL定理表明,对于特定问题,某些算法总是表现更好","NFL定理强调算法的性能与问题的分布无关","NFL定理适用于所有监督学习和非监督学习场景","下列关于NFL定理的说法中,哪一项是正确的","v1",0,{"answer":34,"createTime":5,"id":35,"options":36,"question":41,"source":31,"type":32},[],336173400,[37,38,39,40],"准确率 (Accuracy)","精确率 (Precision)","均方误差 (MSE)","召回率 (Recall)","在模型评估中,以下哪个指标常用于回归问题",{"answer":43,"createTime":5,"id":44,"options":45,"question":50,"source":31,"type":32},[],336173401,[46,47,48,49],"衡量模型的复杂度","量化模型预测值与真实值之间的差异","计算模型的梯度方向","用于生成新的训练样本","在回归模型中,损失函数的主要作用是什么",{"answer":52,"createTime":5,"id":53,"options":54,"question":59,"source":31,"type":32},[],336173402,[55,56,57,58],"\u003Cimg src=\"https:\u002F\u002Ftihai-oss-cloud.itihey.com\u002Fimg\u002F3fe9769614f18517d0ea95da0f751fd4.png\">","\u003Cimg src=\"https:\u002F\u002Ftihai-oss-cloud.itihey.com\u002Fimg\u002F01d9ff21184d7ab4bae6287987d49ae5.png\">","\u003Cimg src=\"https:\u002F\u002Ftihai-oss-cloud.itihey.com\u002Fimg\u002F57a34c34eb517bd17370d4ca5269ae87.png\">","\u003Cimg src=\"https:\u002F\u002Ftihai-oss-cloud.itihey.com\u002Fimg\u002Fd7bac60b563f8ce2cae525f98638b8a8.png\">","在线性回归中,假设函数的形式是",{"answer":61,"createTime":5,"id":62,"options":63,"question":68,"source":31,"type":32},[],336173403,[64,65,66,67],"控制模型的复杂程度","决定模型是否收敛","调节每次参数更新的步长大小","影响模型预测的速度","在优化模型参数时,学习率的作用是什么",{"answer":70,"createTime":5,"id":71,"options":72,"question":77,"source":31,"type":32},[],336173404,[73,74,75,76],"整个训练集","一个样本","一个批次的样本","随机选择的10个样本","在小批量随机梯度下降法中,每次迭代使用的数据量是",{"answer":79,"createTime":5,"id":80,"options":81,"question":86,"source":31,"type":32},[],336173405,[82,83,84,85],"模型训练速度变慢","模型无法收敛甚至发散","参数更新幅度太小","模型更容易过拟合","在模型训练中,若学习率过高,会导致什么问题",{"answer":88,"createTime":5,"id":89,"options":90,"question":96,"source":31,"type":97},[],336173406,[91,92,93,94,95],"偏差指的是模型的系统性错误,方差指的是模型对训练数据的敏感性","偏差高通常意味着模型过于复杂,而方差高通常意味着模型过于简单","偏差和方差之间存在权衡关系,降低其中一个可能会增加另一个","可以通过增加训练样本数量来同时减小偏差和方差","偏差主要影响模型的拟合能力,而方差主要影响其泛化能力","在机器学习中,关于偏差与方差的关系,以下哪些说法是正确的",1,{"answer":99,"createTime":5,"id":100,"options":101,"question":107,"source":31,"type":97},[],336173407,[102,103,104,105,106],"留出法","K折交叉验证","留一法","分层K折交叉验证法","随机采样法","在模型评估中,以下哪些方法可以用于数据集的划分",{"answer":109,"createTime":5,"id":6,"options":110,"question":20,"source":31,"type":97},[],[8,9,10,11,12]]