Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Fair Streaming Feature Selection

Zhangling Duan, Tianci Li, Xingyu Wu, Zhaolong Ling, Jingye Yang, Zhaohong Jia

TL;DR

Empirical evaluations show that FairSFS not only maintains accuracy that is on par with leading streaming feature selection methods and existing fair feature techniques but also significantly improves fairness metrics.

Abstract

Streaming feature selection techniques have become essential in processing real-time data streams, as they facilitate the identification of the most relevant attributes from continuously updating information. Despite their performance, current algorithms to streaming feature selection frequently fall short in managing biases and avoiding discrimination that could be perpetuated by sensitive attributes, potentially leading to unfair outcomes in the resulting models. To address this issue, we propose FairSFS, a novel algorithm for Fair Streaming Feature Selection, to uphold fairness in the feature selection process without compromising the ability to handle data in an online manner. FairSFS adapts to incoming feature vectors by dynamically adjusting the feature set and discerns the correlations between classification attributes and sensitive attributes from this revised set, thereby forestalling the propagation of sensitive data. Empirical evaluations show that FairSFS not only maintains accuracy that is on par with leading streaming feature selection methods and existing fair feature techniques but also significantly improves fairness metrics.

Fair Streaming Feature Selection

TL;DR

Empirical evaluations show that FairSFS not only maintains accuracy that is on par with leading streaming feature selection methods and existing fair feature techniques but also significantly improves fairness metrics.

Abstract

Streaming feature selection techniques have become essential in processing real-time data streams, as they facilitate the identification of the most relevant attributes from continuously updating information. Despite their performance, current algorithms to streaming feature selection frequently fall short in managing biases and avoiding discrimination that could be perpetuated by sensitive attributes, potentially leading to unfair outcomes in the resulting models. To address this issue, we propose FairSFS, a novel algorithm for Fair Streaming Feature Selection, to uphold fairness in the feature selection process without compromising the ability to handle data in an online manner. FairSFS adapts to incoming feature vectors by dynamically adjusting the feature set and discerns the correlations between classification attributes and sensitive attributes from this revised set, thereby forestalling the propagation of sensitive data. Empirical evaluations show that FairSFS not only maintains accuracy that is on par with leading streaming feature selection methods and existing fair feature techniques but also significantly improves fairness metrics.
Paper Structure (13 sections, 5 equations, 10 figures, 7 tables, 1 algorithm)

This paper contains 13 sections, 5 equations, 10 figures, 7 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. Related work
  3. Streaming feature selection
  4. Fair machine learning
  5. Definitions
  6. Fair streaming feature selection
  7. Algorithm implementation
  8. Algorithm analysis
  9. Experiments
  10. Experimental setup
  11. Comparison with streaming feature selection
  12. Comparison with fair feature selection algorithms
  13. Conclusion

Figures (10)

  • Figure 1: The use of gender features in the content recommendation process may lead to bias and discrimination.
  • Figure 2: Flowchart of the FairSFS Algorithm
  • Figure 3: Radar graph depicting the fairness performance of FairSFS alongside its competitors in streaming feature selection, focusing on the metrics SPD (left) and PE (right) when using the KNN classifier(where lower scores for SPD and PE denote increased fairness in the model).
  • Figure 4: Radar graph depicting the fairness performance of FairSFS alongside its competitors in streaming feature selection, focusing on the metrics SPD (left) and PE (right) when using the NB classifier.
  • Figure 5: Radar graph depicting the fairness performance of FairSFS alongside its competitors in streaming feature selection, focusing on the metrics SPD (left) and PE (right) when using the LR classifier.
  • ...and 5 more figures

Theorems & Definitions (3)

  • proof
  • proof
  • proof