Machine Learning

Fundamentals of Machine Learning (FML)

1. Supervised Learning

   • Threshold classifier
   • Nearest Neighbor classifier
   • Nearest Mean classifier
   • Quadratic Discriminant Analysis
   • Linear Discriminant Analysis
   • Density Tree and Decision Tree
   • Density Forest and Decision Forest
   • Linear Regression and Regression Tree
   • Bias-Variance Tradeoff
   • Ridge Regression
   • Automatic Feature Selection for Regression
     • Orthogonal Matching Pursuit
   • Non-linear Regression
     • Gaussian-Newton Algorithm
     • Levenberg–Marquardt Algorithm
   • Kernel Ridge Regression
   • Robust Regression: RANSAC

2. Unsupervised Learning

   • Non-negative matrix factorization

3. Feature and model selection

   • K-Fold Cross Validation

4. Extras

   • Inverse Transform Sampling:

     For generative models, one can create data using a random number generator. Specifically, one first samples an instance label Y according to the prior probabilities, and then uses the corresponding likelihood to sample the feature X. If no predefined random generator for the desired likelihood is available, uniformly distributed samples from a standard random number generator can be transformed to the desired distribution by means of 'inverse transform sampling' (see reference)

   • Principal Component Analysis (PCA)

   • Singular Value Decomposition (SVD)

   • Recommender System

Advanced Machine Learning (FML)