machine learning code for Northeastern University CS6140 fall 2015
(1) Some models use Java multi-thread to optimize speed
- Regression Tree & Classification Tree
- Linear Regression
- Logistic Regression
- Perceptron
- Neural Network
- C-SVMs:
1) Linear kernel
2) Polynomial kernel
3) Gaussian kernel
- Gaussian Discriminant Analysis
1) Single Gaussian Model
2) Mixed Gaussian Model
- Naive Bayes:
1) Bernoulli + Multinoulli
2) Gaussian
- Boosting:
1) AdaBoost by classification tree for classification
2) GradientBoost by regression tree for regression and classification
- ECOC:
1) On top of AdaBoost
2) On top of SVMs
- Bagging / RandomForest on top of Tree
- KNN with various kernels
- PCA
(2) Implemented Gradient Decent framework to multi-thread update especially for neural network
(3) Designed and implemented Neural Network for any desired structure with NLL objective
(4) Implemented Cross-Validation process.
(5) Implemented two type of feature matrix:
- FullMatrix
- SparseMatrix