Sklearn is a popular machine-learning library (https://github.com/scikit-learn/scikit-learn). It has a large code base with hundreds of thousands of lines of code. The primary target of this project is to make sklearn-mini with many important functionalities of sklearn, and remain under 1000 lines of code.
This would make the library very usable and readable.
- Models
- Regression
- Linear regression
- Regression Trees
- MLP
- Classification
- Logistic Regression
- Classification Trees
- MLP
- SVM
- k-nearest neighbours
- Dimensionality reduction
- PCA?
- Clustering
- K-means
- Regression
- Datasets
- Load some basic datasets like titanic, iris, etc.
- Ability to make datasets
- Model_Selection
- Select best models' weights, hyperparameters, etc.
- Preprocessing
- Train-test split
- Built on Numpy
- Probably no Cuda support (because of 1K Loc limit)