-
error rate:错误率
-
accuracy:精度
-
error:误差
-
training error:训练误差(或“empirical error:经验误差”)
-
test error:测试误差
-
generalization error:泛化误差
-
overfitting:过拟合
-
underfitting:欠拟合
-
model selection:模型选择
-
hold-out:留出法
-
sampling:采样
-
stratified sampling:分层采样
-
cross validation:交叉验证
-
k-fold cross validation:k折交叉验证
-
validation set:验证集
-
Leave-One-Out:留一法(简称“LOO”)
-
bootstrapping:自助法
-
bootstrap sampling:自助采样
-
out-of-bag estimate:包外估计
-
parameter tunning:调参
-
performance measure:性能度量
-
mean squared error:均方误差
-
precision:查准率,亦称“准确率”
-
recall:查全率,亦称“召回率”
-
true positive:真正例
-
True Positive Rate:真正例率(简称"TPR")
-
false positive:假正例
-
False Positive Rate:假正例率(简称"FPR")
-
true negative:真反例
-
false negative:假反例
-
confusion matrix:混淆矩阵
-
Break-Evendors Point:平衡点(简称“BEP”)
-
threshold:阈值
-
cut point:截断点
-
ROC:受试者工作曲线
-
AUC:Area under ROC Curve
-
loss:损失
-
unequal cost:非均等代价
-
cost-sensitive:代价敏感
-
hypothesis test:假设检验
-
binomial test:二项测试
-
confidence:置信度
-
two-tailed:双边
-
t-test:t检验
-
paired t-test:成对t检验
-
contingency table:列联表
-
bias-variance decomposition:偏差-方差分解
-
bias-variance dilemma:偏差-方差窘境
-
cost-sensitive learning:代价敏感学习
-
harmonic mean:调和平均
《机器学习》基本术语标准翻译总结:第2章
最新推荐文章于 2022-08-02 19:45:28 发布