ridge
Python Machine Learning Library specialized in L2R and Recommendation.
Named after my favorite racing game R4.
日本語のREADMEはこちら．
Requirements
- Python 3.6.0 ~
- NumPy
- Scipy
Implementation
These models are available.
| Class | Module | Description | Document |
|---|---|---|---|
| FMRegressor | factorization_machines | Factorization Machine for regression tasks | FMs |
| FMClassifier | factorization_machines | Factorization Machine for classification tasks | FMs |
| MatFac | matrix_factorization | Ordinal Matrix Factorization | MFs |
| NNMatFac | matrix_factorization | Non-negative Matrix Factorization | MFs |
| ConditionalLogit | logit_models | Conditional Logistic Regression as a discrete choice model | LMs |
Directories
| name | descripiton |
|---|---|
| ridge | Model implementation |
| ridge.racer | Cython optimization |
| docs | Documents (usage of this package & memo related to implementation) |
| tests | Model test suites |
Models
MFs
MatFac (Matrix Factorization)
- MovieLens 100k
- with Cython 17m20s (1000epochs)
- with Python 28m38s (1000epochs)