Cost-Sensitive Multi-Class Active Learning
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Failed to load latest commit information.
query_strategies
vehicle
.gitignore
README.md
demo.py
result.png

README.md

Cost-Sensitive Multi-Class Active Learning

Python implementation of our paper A Novel Uncertainty Sampling Algorithm for Cost-Sensitive Multiclass Active Learning and related algorithms, including:

  • Active Learning with Cost Embedding (ALCE)
  • Uncertainty Sampling with Margin (UncertaintyMargin)
  • Uncertainty Sampling with Entropy (UncertaintyEntropy)
  • Maximum Expected Cost (MEC)
  • Cost-Weighted Minimum Margin (CWMM)
  • Random Sampling

If you find our paper or implementation is useful in your research, please consider citing our paper for ALCE and the references below for other algorithms.

@inproceedings{Huang2016alce,
    author    = {Kuan-Hao Huang and
                 Hsuan-Tien Lin},
    title     = {A Novel Uncertainty Sampling Algorithm for Cost-Sensitive Multiclass Active Learning},
    booktitle = {Proceedings of the IEEE International Conference on Data Mining (ICDM)},
    pages     = {925--930},
    year      = {2016},
}

Prerequisites

  • Python 2.7.10
  • NumPy 1.9.2
  • scikit-learn 0.19.1
  • Matplotlib 1.3.1

Usage

$ python demo.py

Dataset

Result

result.png

Reference

  • Simon Tong and Daphne Koller. Support Vector Machine Active Learning with Applications to Text Classification. Journal of Machine Learning Research, 2001

  • Feng Jing, Mingjing Li, HongJiang Zhang, and Bo Zhang. Entropy-Based Active Learning with Support Vector Machines for Content-Based Image Retrieval. ICME, 2004

  • Po-Lung Chen and Hsuan-Tien Lin. Active Learning for Multiclass Cost-Sensitive Classification Using Probabilistic Models. TAAI, 2013

  • Kuan-Hao Huang and Hsuan-Tien Lin. A Novel Uncertainty Sampling Algorithm for Cost-Sensitive Multiclass Active Learning. ICDM, 2016

Author

Kuan-Hao Huang / @ej0cl6