A personal machine learning library in Python Anything involving deep learning is over at https://github.com/rohan-varma/neuralnets.
Algorithms with linear decision boundaries (perceptrons, logistic regression, linear regression) are in the linear_algorithms folder, algorithms with non-linear decision boundaries (k-nearest neighbors, decision trees) are at the nonlinear_algorithms directory.
Currently Implemented:
- Basic sampling from distributions and plotting them on matplotlib
- Perceptron algorithm
- Voting perceptron
- K-Fold Cross validation to detect overfitting
- Hyperparameter tuning with K-Fold CV, and other stuff
- Logistic Regression
- Linear Regression
Todo:
- Support for feature transformations
- SVMs
- Decision Trees
- Kernel functions
- Multiclass logistic regression
- Multiclass SVMs
- k nearest neighbors
- ridge regression
- lasso regression
- kernelized regression
- kernelized nearest neighbors
- kernelized SVM