https://github.com/lorey/incremental-learning-with-scikit-learn
Documenting incremental learning with partial_fit
https://github.com/lorey/incremental-learning-with-scikit-learn
Last synced: 7 months ago
JSON representation
Documenting incremental learning with partial_fit
- Host: GitHub
- URL: https://github.com/lorey/incremental-learning-with-scikit-learn
- Owner: lorey
- Created: 2023-06-07T23:00:27.000Z (almost 3 years ago)
- Default Branch: master
- Last Pushed: 2024-06-18T00:31:10.000Z (almost 2 years ago)
- Last Synced: 2025-03-15T08:55:05.499Z (about 1 year ago)
- Language: Python
- Size: 8.79 KB
- Stars: 0
- Watchers: 2
- Forks: 1
- Open Issues: 2
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# Incremental Learning with scikit-learn
This repo implements incremental learning with scikit-learn
by using a sample dataset containing a concept drift,
i.e. a target variable that changes its meaning over time.
### Scenario
Users choose items they like. While they reject everything in the beginning `f(g)=0`,
a regular pattern emerges over time `f(g)=-x`.
I used this sample to understand the effects of `partial_fit`
with respect to both fitting time and quality of the training.
### Methods
- `partial_all`: partial_fit on all data until this point
- `partial_step`: partial_fit on batch data
- `full_step`: regular fit on batch data (ignores all previous samples!)
- `full_all`: regular fit on all data until this point
PS: Please be aware that I have no idea what I'm doing.