Feature and Instance Joint Selection: A Reinforcement Learning Perspective

@inproceedings{Fan2022FeatureAI,
  title={Feature and Instance Joint Selection: A Reinforcement Learning Perspective},
  author={Wei Fan and Kunpeng Liu and Hao Liu and Hengshu Zhu and Hui Xiong and Yanjie Fu},
  booktitle={IJCAI},
  year={2022}
}
Feature selection and instance selection are two important techniques of data processing. However, such selections have mostly been studied separately, while existing work towards the joint selection conducts feature/instance selection coarsely; thus neglecting the latent fine-grained interaction between feature space and instance space. To address this challenge, we propose a reinforcement learning solution to accomplish the joint selection task and simultaneously capture the interaction… 

Figures and Tables from this paper

References

SHOWING 1-10 OF 42 REFERENCES
Automating Feature Subspace Exploration via Multi-Agent Reinforcement Learning
TLDR
A multi-agent reinforcement learning framework for the feature selection problem that could search the feature subset space globally and could be easily adapted to the real-time case (real-time feature selection) due to the nature of reinforcement learning.
AutoFS: Automated Feature Selection via Diversity-aware Interactive Reinforcement Learning
TLDR
An Interactive Reinforced Feature Selection (IRFS) framework is proposed that guides agents by not just self-exploration experience, but also diverse external skilled trainers to accelerate learning for feature exploration and help agents to learn broader knowledge, and thereafter be more effective.
Interactive Reinforcement Learning for Feature Selection with Decision Tree in the Loop
TLDR
A novel interactive and closed-loop architecture to simultaneously model interactive reinforcement learning (IRL) and decision tree feedback (DTF) and a new embedding method capable of empowering graph convolutional network to jointly learn state representation from both the graph and the tree is proposed.
Simplifying Reinforced Feature Selection via Restructured Choice Strategy of Single Agent
TLDR
A single-agent reinforced feature selection approach integrated with restructured choice strategy is developed, which exploits only one single agent to handle the selection task of multiple features, instead of using multiple agents.
Integrating Instance Selection, Instance Weighting, and Feature Weighting for Nearest Neighbor Classifiers by Coevolutionary Algorithms
TLDR
A novel approach integrating instance selection, instance weighting, and feature weighting into the framework of a coevolutionary model is presented, and it is shown that it outperforms other methods in the comparison, thus becoming a suitable tool in the task of enhancing the nearest neighbor classifier.
sCOs: Semi-Supervised Co-Selection by a Similarity Preserving Approach
TLDR
A unified framework, called sCOs, is proposed, which efficiently integrates labeled and unlabeled parts into the co-selection process of semi-supervised learning tasks, and is based on introducing both a sparse regularization term and a similarity preserving approach.
...
...