[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"$fpjUnPskXl6VKjPLK7fnidxCiIgdRPGL8SR8_jDX7WfU":3},{"answer":4,"createTime":5,"id":6,"options":7,"origin":10,"question":17,"related":18,"source":28,"type":110},[],"2025-12-01 16:09:35",246901732,[8,9],"正确","错误",{"count":11,"courseId":12,"courseImg":13,"courseName":14,"workId":15,"workName":16},84,"53e1d2ef4961cca8eea3e23969ad2cb9","https:\u002F\u002Ftihai-oss-cloud.itihey.com\u002Fimg\u002F03a579384a6dc297c89809b582fcc767.png","默认课程","exam_168281068","机器学习初步","决策树剪枝的基本策略只有"后剪枝"",[19,30,39,48,57,66,75,84,93,102],{"answer":20,"createTime":5,"id":21,"options":22,"question":27,"source":28,"type":29},[],246901652,[23,24,25,26],"极大似然估计属于贝叶斯主义","极大似然估计需先假设某种概率分布形式","极大似然估计的任务是利用训练集估计参数","极大似然估计做了独立同分布假设","下列说法错误的是()","v1",0,{"answer":31,"createTime":5,"id":32,"options":33,"question":38,"source":28,"type":29},[],246901653,[34,35,36,37],"指数函数","绝对值函数","对数函数","二次函数","对数线性回归是令广义线性模型中的联系函数为什么函数的特例",{"answer":40,"createTime":5,"id":41,"options":42,"question":47,"source":28,"type":29},[],246901654,[43,44,45,46],"1979","1966","1993","1956","第一篇关于决策树的工作在哪一年发表的",{"answer":49,"createTime":5,"id":50,"options":51,"question":56,"source":28,"type":29},[],246901655,[52,53,54,55],"Gain(D,a)+IV(a)","Gain(D,a)\u002FIV(a)","Gain(D,a)-IV(a)","Gain(D,a)*IV(a)","增益率的表达式是Gain_ratio(D,a)=()",{"answer":58,"createTime":5,"id":59,"options":60,"question":65,"source":28,"type":29},[],246901656,[61,62,63,64],"验证集;需要","验证集;不需要","训练集;不需要","训练集;需要","我们通常将数据集划分为训练集,验证集和测试集进行模型的训练,参数的验证需要在()上进行,参数确定后()重新训练模型",{"answer":67,"createTime":5,"id":68,"options":69,"question":74,"source":28,"type":29},[],246901657,[70,71,72,73],"降低;降低","增加;降低","降低;增加","增加;增加",""一对一"相较于"一对其余",存储开销(),测试时间开销()",{"answer":76,"createTime":5,"id":77,"options":78,"question":83,"source":28,"type":29},[],246901658,[79,80,81,82],"以很低概率得到很好的模型","以很低概率得到不好的模型","以很高概率得到很好的模型","以很高概率得到不好的模型","以下哪个是对概率近似正确(PAC)的正确解释",{"answer":85,"createTime":5,"id":86,"options":87,"question":92,"source":28,"type":29},[],246901659,[88,89,90,91],"对每个样本 x 选择能使后验概率 P(c∣x) 最大的类别标记","对每个样本 x 选择能使条件风险 R(c_i∣x) 最大的类别标记","对每个样本 x 选择能使条件风险 R(c_i∣x)最小的类别标记","对每个样本 x 选择能使后验概率 P(c∣x) 最小的类别标记","以下哪个选项是对贝叶斯最优分类器的描述",{"answer":94,"createTime":5,"id":95,"options":96,"question":101,"source":28,"type":29},[],246901660,[97,98,99,100],"测试集","训练集","验证集","以上选项都可以","调参以什么集合上的性能作为评价标准",{"answer":103,"createTime":5,"id":104,"options":105,"question":109,"source":28,"type":29},[],246901661,[106,107,108],"层次聚类","密度聚类","原型聚类","k 均值聚类属于下列哪一种聚类算法",3]