(1)
函数关系：functional relation
正相关：positive correlation
负相关：negative correlation
相关系数：correlation efficient
一元线性回归：simple linear regression
多元线性回归：multiple linear regression
参数：parameter
参数估计：parameter estimation
截距：intercept
斜率：slope
误差：error
残差：residual
拟合：fit
最小二乘法：method of least squares
残差平方和：residual sum of squares (RSS)
创建向量：create a vector
回归系数：regression coefficient
建立线性模型：create a linear model
内推插值：interpolate
外推归纳：extrapolate
回归诊断：regression diagnostics
离群值：outlier
多重共线性：multicollinearity
广义线性回归模型：generalized linear regression model
哑变量（虚拟变量）：dummy variable
logistic回归：logistic regression
非线性回归：nonlinear regression
(2)
频繁模式挖掘：frequent pattern mining
无偏估计：unbiased estimation
最小二乘法：least square methods (LSM)
矩阵：matrix
系数：coefficient
截距项：intercept
随机误差：random error
广义逆：generalized inverse
有偏估计：biased estimation
岭回归：ridge regression
等价模型：equivalence model
惩罚模型：penalty function
估计族：class of estimators
岭参数：ridge parameter
岭迹图：ridge trace
方差扩大因子：variance inflation factor (VIF)
LASSO：least absolute shrinkage and selection operator
最小角回归：least angle regression (LAR)
(3)
主成分分析：principal component analysis
降维：dimension reduction
特征选择：feature selection
特征提取：feature extraction
协方差矩阵：covariance matrix
对角化：diagonalization
因子分析：factor analysis
因子载荷矩阵：factor loading matrix
特殊方差矩阵：special variance matrix
极大似然法：maximum likelihood method
正交旋转：orthorgonal rotation
因子得分：factor score
(4)
分类模型：classification model
线性判别法：linear discriminant analysis
线性分类器：Linear Classifier
距离判别法：distance discriminant method
k-NN algorithm（k-最近邻算法）：k-Nearest Neighbors algorithm (or k-NN for short)
朴素贝叶斯分类器：Naïve Bayes Classifier
决策树：decision tree
支持向量机：support vector machines (SVM)
神经网络：neural network
文本挖掘：text mining
贝叶斯信念网络：Bayesian belief network
条件概率：conditional probability
先验概率：prior probability
后验概率：posterior probability
条件概率表：conditional probability table (CPT)
贝叶斯推理：Bayesian reasoning
(5)
决策树：decision tree
ID3算法（迭代二叉树3代）：iterative dichotomiser 3
分类与回归树：classification and regression tree(CART)
信息增益：information gain (IG)
分裂属性：splitting attribute
剪枝：pruning
代价复杂度：cost-complexity
组合算法：combinatorial algorithm
装袋算法：bagging algorithm
提升算法：boosting algorithm
AdaBoost（Adaptive Boosting：
自适应增强）算法：Adaboost algorithm
随机森林算法：random forest algorithm
(6)
支持向量机：support vector machines (SVM)
线性可分：linearly-separable
最优分离平面：optimal separating plane
决策边界：decision boundary
最大边缘超平面：maximum marginal hyperplane (MMH)
凸优化问题：convex optimization problem
拉格朗日乘子法：lagrange multiplier method
KKT条件：Karush-Kuhn-Tucker conditions
对偶：duality
松弛变量：slack variable
惩罚函数：penalty function
SMO算法（序列最小优化算法）：sequential minimal optimization
高维空间：high dimension space
维度灾难：curse of dimensionality
核函数：kernel function
高斯径向基函数：Gaussian radial basis function
Mercer定理：Mercer's theorem
(7)
人工神经网络：artificial neural network
神经元，神经细胞：neuron
树突：dendrite
轴突：axon
细胞体：cell body
突触：synapsis
单层感知器：single-layer perceptron
线性神经网络：linear neural network
BP神经网络：BP Neural Network
输入节点：input node
输出节点：output node
权向量：weight vector
偏置因子（偏因）：bias factor
激活函数：activation function
自学习算法：self-learning algorithm
学习率：learning rate
权重：weight
偏移：bias
偏置值：bias value
平均绝对误差：mean absolute error
均方误差：mean square error
误差平方和：square error sum
拓扑：topology
前馈型网络：feedforward networks
反馈型网络：feedback network
学习规则：learning rule
线性神经网络：linear neural network
梯度下降法：gradient descent algorithm
(8)
深度学习：deep learning
有监督学习：supervised learning
无监督学习：unsupervised learning
半监督学习：semi-supervised learning
BP神经网络：BP Neural Network
误差反向传播算法：error back-propagation algorithm
多层前馈神经网络：multilayer feed-forward neural network
最速下降法：method of steepest descent
图像压缩：image compression
Hopfield神经网络：Hopfield neural network
光学字符识别：Optical Character Recognition (OCR)
PCA（Principal Component Analysis）神经网络：PCA Neural Network
神经网络芯片：neural network chip
通用逼近器：universal approximator
径向基函数神经网络：radial basis function neural network (RBFNN)
正则化RBF神经网络：normalized RBF neural network (NRBFNN)）
广义RBF神经网络：Generalized RBF neural network
概率神经网络：probabilistic neural network
贝叶斯信念网络：Bayesian belief network
(9)
贝叶斯信念网络：Bayesian belief network
梯度计算：gradient computation
权重更新：weight updating
聚类：clustering
孤立点（离群值、异常值）：outlier
距离：distance
绝对值距离：absolute value distance
欧氏（欧几里得）距离：Euclidean distance
闵可夫斯基：Minkowski distance
切比雪夫距离：Chebyshev distance
马氏（马哈拉诺比斯）距离：Mahalanobis distance
兰氏距离：Lance and Williams distance / Canberra Distance
离散变量：discrete variable
数据中心化与标准化：data centralization and standardization
极差（全距）：range
相似系数：similarity coefficient
夹角余弦：included angle cosine
相关系数：correlation coefficient
凝聚聚类算法：aggregate clustering algorithm
层次聚类法：hierarchical clustering
动态聚类：dynamic clustering
K-means 算法：K-means algorithm
K中心聚类法：k-medoids clustering
围绕中心点的划分算法：Partitioning Around Medoids (PAM)
CLARA 算法：CLARA algorithm (Clustering for LARge Applications)
基于密度的聚类：density-based clustering
DBSCAN 聚类算法：DBSCAN (A Density-Based Spatial Clustering of Application) clustering algorithm
基于网格的聚类算法：Grid-Based Clustering Algorithm
CLIQUE算法：CLIQUE (CLustering In QUEst) Algorithm
稠密单元：dense unit
簇的装配：assembling of cluster
最小覆盖：minimum covering
贪心算法：greedy algorithm
(10)
霍普金斯统计量：Hopkins statistic
簇数制定：determining the number of clusters
肘方法：elbow method
伪F统计量：pseudo f-statistics (PSF)
伪T平方统计量：Pseudo-T2 Statistic (PST2)：
B-cubed聚类评分：B-cubed cluster scoring
轮廓系数：silhouette coefficient
模糊聚类：fuzzy clustering
误差平方和：square sum of error (SSE)
基于概率模型的聚类：probabilistic-model-based clustering
概率聚类：probabilistic clustering
最大似然估计：maximum likelihood estimation
最大期望算法：expectation-maximization (EM) algorithm
混合模型：hybrid model
离群值检测：outlier detection
基于直方图的离群值检测：histogram-based outlier detection
基于邻域的离群值检测：neighborhood-based outlier detection
基于网格的离群值检测：grid-based outlier detection
基于聚类的离群值检测：clustering-based outlier detection
posted on
