03 Text Classification - Mind Map

Up: 03 Text Classification - Index | 03 Text Classification - Cheat Sheet

mindmap
  root((Text Classification))
    Probability pre-study
      Conditional probability
      Density estimation
      Argmax
      MLE
      Bayes rule
      MAP
    Naive Bayes
      Conditional independence
      Mega-document estimate
      Zero probability / OOV
      Laplace smoothing
      Multinomial NB
      Bernoulli NB
      Gaussian NB
      Log trick / underflow
    Logistic regression
      Sigmoid
      Cross-entropy loss
      Gradient descent
      Decision boundary
      One-vs-all
      Softmax
    Evaluation
      Confusion matrix
      Precision / Recall
      F1
      ROC / AUC
      Cross validation
    Decision trees
      ID3
      Entropy
      Information gain
      Gain ratio
      Gini
      Pruning
    Generalization
      Overfitting
      Underfitting
      Bias-variance
    Ensembles
      Bagging
      Random forest
      Boosting
      Stacking
      No free lunch
    Applications
      Sentiment analysis
      Word sense disambiguation
      Labels and agreement

How the pieces depend on each other

flowchart TD
    B["Bayes rule + MAP"] --> NB["Naive Bayes<br/>(cond. independence)"]
    NB --> S["Smoothing<br/>(zero prob. / OOV)"]
    NB --> V["MNB / BNB / GNB"]
    V --> L["Log trick<br/>(underflow)"]
    NB -. "generative vs discriminative" .-> LR["Logistic regression<br/>(sigmoid, CE loss, GD)"]
    LR --> SM["Softmax / one-vs-all"]
    NB --> E["Evaluation<br/>P, R, F1, ROC"]
    LR --> E
    DT["Decision trees<br/>(entropy, IG, Gini)"] --> OF["Overfitting + pruning"]
    OF --> EN["Ensembles<br/>bagging, boosting, stacking"]
    E --> EN
    EN --> NFL["No free lunch"]