Basis(基礎):
MSE(Mean Square Error 均方誤差),LMS(LeastMean Square 最小均方),LSM(Least Square Methods 最小二乘法),MLE(MaximumLikelihood Estimation最大似然估計),QP(Quadratic Programming 二次規劃), CP(Conditional Probability條件概率),JP(Joint Probability 聯合概率),MP(Marginal Probability邊緣概率),Bayesian Formula(貝葉斯公式),L1 /L2Regularization(L1/L2正則,以及更多的,現在比較火的L2.5正則等),GD(GradientDescent 梯度下降),SGD(Stochastic Gradient Descent 隨機梯度下降),Eigenvalue(特徵值),Eigenvector(特徵向量),QR-decomposition(QR分解),Quantile (分位數),Covariance(協方差矩陣)。
Common Distribution(常見分佈):
Discrete Distribution(離散型分佈):BernoulliDistribution/Binomial(貝努利分佈/二項分佈),Negative BinomialDistribution(負二項分佈),MultinomialDistribution(多項式分佈),Geometric Distribution(幾何分佈),HypergeometricDistribution(超幾何分佈),Poisson Distribution (泊松分佈)
Continuous Distribution (連續型分佈):UniformDistribution(均勻分佈),Normal Distribution /Guassian Distribution(正態分佈/高斯分佈),ExponentialDistribution(指數分佈),Lognormal Distribution(對數正態分佈),GammaDistribution(Gamma分佈),Beta Distribution(Beta分佈),Dirichlet Distribution(狄利克雷分佈),Rayleigh Distribution(瑞利分佈),Cauchy Distribution(柯西分佈),Weibull Distribution (韋伯分佈)
Three Sampling Distribution(三大抽樣分佈):Chi-squareDistribution(卡方分佈),t-distribution(t-distribution),F-distribution(F-分佈)
Data Pre-processing(數據預處理):
Missing Value Imputation(缺失值填充),Discretization(離散化),Mapping(映射),Normalization(歸一化/標準化)。
Sampling(採樣):
Simple Random Sampling(簡單隨機採樣),OfflineSampling(離線等可能K採樣),Online Sampling(在線等可能K採樣),Ratio-based Sampling(等比例隨機採樣),Acceptance-RejectionSampling(接受-拒絕採樣),Importance Sampling(重要性採樣),MCMC(MarkovChain Monte Carlo 馬爾科夫蒙特卡羅採樣算法:Metropolis-Hasting& Gibbs)。
Clustering(聚類):
K-Means,K-Mediods,二分K-Means,FK-Means,Canopy,Spectral-KMeans(譜聚類),GMM-EM(混合高斯模型-期望最大化算法解決),K-Pototypes,CLARANS(基於劃分),BIRCH(基於層次),CURE(基於層次),DBSCAN(基於密度),CLIQUE(基於密度和基於網格)
Classification&Regression(分類&迴歸):
LR(Linear Regression 線性迴歸),LR(LogisticRegression邏輯迴歸),SR(Softmax Regression 多分類邏輯迴歸),GLM(GeneralizedLinear Model 廣義線性模型),RR(Ridge Regression 嶺迴歸/L2正則最小二乘迴歸),LASSO(Least Absolute Shrinkage andSelectionator Operator L1正則最小二乘迴歸), RF(隨機森林),DT(DecisionTree決策樹),GBDT(Gradient BoostingDecision Tree 梯度下降決策樹),CART(ClassificationAnd Regression Tree 分類迴歸樹),KNN(K-Nearest Neighbor K近鄰),SVM(Support VectorMachine),KF(KernelFunction 核函數PolynomialKernel Function 多項式核函數、Guassian KernelFunction 高斯核函數/Radial BasisFunction RBF徑向基函數、String KernelFunction 字符串核函數)、 NB(Naive Bayes 樸素貝葉斯),BN(Bayesian Network/Bayesian Belief Network/ Belief Network 貝葉斯網絡/貝葉斯信度網絡/信念網絡),LDA(Linear Discriminant Analysis/FisherLinear Discriminant 線性判別分析/Fisher線性判別),EL(Ensemble Learning集成學習Boosting,Bagging,Stacking),AdaBoost(Adaptive Boosting 自適應增強),MEM(MaximumEntropy Model最大熵模型)
Effectiveness Evaluation(分類效果評估):
Confusion Matrix(混淆矩陣),Precision(精確度),Recall(召回率),Accuracy(準確率),F-score(F得分),ROC Curve(ROC曲線),AUC(AUC面積),LiftCurve(Lift曲線) ,KS Curve(KS曲線)。
PGM(Probabilistic Graphical Models概率圖模型):
BN(Bayesian Network/Bayesian Belief Network/ BeliefNetwork 貝葉斯網絡/貝葉斯信度網絡/信念網絡),MC(Markov Chain 馬爾科夫鏈),HMM(HiddenMarkov Model 馬爾科夫模型),MEMM(Maximum Entropy Markov Model 最大熵馬爾科夫模型),CRF(ConditionalRandom Field 條件隨機場),MRF(MarkovRandom Field 馬爾科夫隨機場)。
NN(Neural Network神經網絡):
ANN(Artificial Neural Network 人工神經網絡),BP(Error BackPropagation 誤差反向傳播)
Deep Learning(深度學習):
Auto-encoder(自動編碼器),SAE(Stacked Auto-encoders堆疊自動編碼器:Sparse Auto-encoders稀疏自動編碼器、Denoising Auto-encoders去噪自動編碼器、Contractive Auto-encoders 收縮自動編碼器),RBM(RestrictedBoltzmann Machine 受限玻爾茲曼機),DBN(Deep Belief Network 深度信念網絡),CNN(ConvolutionalNeural Network 卷積神經網絡),Word2Vec(詞向量學習模型)。
DimensionalityReduction(降維):
LDA LinearDiscriminant Analysis/Fisher Linear Discriminant 線性判別分析/Fisher線性判別,PCA(Principal Component Analysis 主成分分析),ICA(IndependentComponent Analysis 獨立成分分析),SVD(Singular Value Decomposition 奇異值分解),FA(FactorAnalysis 因子分析法)。
Text Mining(文本挖掘):
VSM(Vector Space Model向量空間模型),Word2Vec(詞向量學習模型),TF(Term Frequency詞頻),TF-IDF(Term Frequency-Inverse DocumentFrequency 詞頻-逆向文檔頻率),MI(MutualInformation 互信息),ECE(Expected Cross Entropy 期望交叉熵),QEMI(二次信息熵),IG(InformationGain 信息增益),IGR(Information Gain Ratio 信息增益率),Gini(基尼係數),x2 Statistic(x2統計量),TEW(TextEvidence Weight文本證據權),OR(Odds Ratio 優勢率),N-Gram Model,LSA(Latent Semantic Analysis 潛在語義分析),PLSA(ProbabilisticLatent Semantic Analysis 基於概率的潛在語義分析),LDA(Latent DirichletAllocation 潛在狄利克雷模型)
Association Mining(關聯挖掘):
Apriori,FP-growth(Frequency Pattern Tree Growth 頻繁模式樹生長算法),AprioriAll,Spade。
Recommendation Engine(推薦引擎):
DBR(Demographic-based Recommendation 基於人口統計學的推薦),CBR(Context-basedRecommendation 基於內容的推薦),CF(Collaborative Filtering協同過濾),UCF(User-basedCollaborative Filtering Recommendation 基於用戶的協同過濾推薦),ICF(Item-basedCollaborative Filtering Recommendation 基於項目的協同過濾推薦)。
Similarity Measure&Distance Measure(相似性與距離度量):
Euclidean Distance(歐式距離),ManhattanDistance(曼哈頓距離),Chebyshev Distance(切比雪夫距離),MinkowskiDistance(閔可夫斯基距離),Standardized Euclidean Distance(標準化歐氏距離),MahalanobisDistance(馬氏距離),Cos(Cosine 餘弦),HammingDistance/Edit Distance(漢明距離/編輯距離),JaccardDistance(傑卡德距離),Correlation Coefficient Distance(相關係數距離),InformationEntropy(信息熵),KL(Kullback-Leibler Divergence KL散度/Relative Entropy 相對熵)。
Optimization(最優化):
Non-constrainedOptimization(無約束優化):Cyclic VariableMethods(變量輪換法),Pattern Search Methods(模式搜索法),VariableSimplex Methods(可變單純形法),Gradient Descent Methods(梯度下降法),Newton Methods(牛頓法),Quasi-NewtonMethods(擬牛頓法),Conjugate Gradient Methods(共軛梯度法)。
ConstrainedOptimization(有約束優化):Approximation Programming Methods(近似規劃法),FeasibleDirection Methods(可行方向法),Penalty Function Methods(罰函數法),Multiplier Methods(乘子法)。
Heuristic Algorithm(啓發式算法),SA(SimulatedAnnealing,模擬退火算法),GA(genetic algorithm遺傳算法)
Feature Selection(特徵選擇算法):
Mutual Information(互信息),DocumentFrequence(文檔頻率),Information Gain(信息增益),Chi-squared Test(卡方檢驗),Gini(基尼係數)。
Outlier Detection(異常點檢測算法):
Statistic-based(基於統計),Distance-based(基於距離),Density-based(基於密度),Clustering-based(基於聚類)。
Learning to Rank(基於學習的排序):
Pointwise:McRank;
Pairwise:RankingSVM,RankNet,Frank,RankBoost;
Listwise:AdaRank,SoftRank,LamdaMART;
Tool(工具):
MPI,Hadoop生態圈,Spark,BSP,Weka,Mahout,Scikit-learn,PyBrain…