mlp_lesson1

Introductino to Scikit-learn

Ch1 Introduction

• Scikit-learn focuses on modelling data
  • Datasets
  • Supervised models
  • Feature selection
  • Clustering of unlabelled data
  • Dataset transformation
    • Preprocessing
    • Feature extraction
    • Normalisation
    • Dimensionality reduction
  • Model selection and evaluation
    • Cross validation
    • Hyperparameter tuning
    • Classification metrics
  • Ensemble methods
    • Boosting
    • Bagging
    • Random Forest

Ch2 Classification Pipeline

• Arrange data into X and y
• Choose the model
• Initialise your model with some hyperparameters
• Fit model to X and y
• Predict labels y_hat_test for X_test
• Evaluate model performance by comparing y_hat_test against y_test

Ch5 Preprocessing

• preprocessing.scale(x)
  • Scale to zero mean and unit variance
• preprocessing.MinMaxScalar().fit_transform(x)
  • Scale values from 0 to 1

Ch6 Classification Models

• kNN
  • sklearn.neighbors - KNeighborClassifier
• Decision Tree
  • sklearn.tree - DecisionTreeClassifier
• Logistic Regression
  • sklearn.linear_model - LogisiticRegression
• Random Forests
  • sklearn.ensemble - RandomForestClassifier
• Naive Bayes (Gaussian)
  • sklearn.ensemble - GaussianNB
• Support Vector Machine
  • sklearn.svm - SVC
• Multiplayer Perceptron (Neural Nets)
  • sklearn.neural_network - MLPClassifier