Implementation of basic ML models and some demo applications
implementaion of basic and most popular machine learning models, including:
Classifier
- k nearest neighbors
- logistic regression: gradient descent
- decision tree: ID3 algorithm
- naive bayes
- SVM: SMO
Predictor
- linear regression: normal equation
Cluster
- k-means
- GMM: EM algorithm
Pre-processor
- standardization
- normalization: L1 norm, L2 norm
- Dimension reduction: PCA, Exhaustive-search selection
- Encoder: one-hot encoding
Model Selection
- Error estimator: resubstitutions, leave-one-out cross-validation
HMM
- Probability estimation: forward-backward algorithm
- Optiaml hidden path: Viterbi algorithm
- Parameter estimation: Baum-Walch algorithm
Digit Recognition
- Classify handwriten digits using this toolkit.
Restuarant Revenue Prediction
- A kaggle compitition from TFI, more.
- Predict annual sales based on the provided data.
Gene-expression classification
- Search for gene feature sets that best discriminate the two prognosis classes
- These data come from a well-known microarray-based cancer classification study, paper here.