.png){kind=icon}
.png){kind=icon}
Return
Evolutionary multistage multitasking method for feature selection in imbalanced data
.png){kind=icon}
PRE
AI
DOI：10.1016/j.swevo.2024.101821.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
In the domain of machine learning, feature selection plays a pivotal role in enhancing model performance, especially for imbalanced datasets, where traditional methods often fall short. Imbalanced datasets are characterized by a disproportionate distribution of class labels, leading to models that maybe biased towards the majority class and perform poorly on the minority class. Effective feature selection is crucial in addressing this imbalance, as it helps in identifying the most relevant features that contribute to the predictive power of the model. This paper introduces a multistage feature selection algorithm that addresses the issue of data imbalance and optimizes feature selection processes. The first stage integrates the Synthetic Minority Oversampling Technique (SMOTE) with Particle Swarm Optimization (PSO) to explore the interactions between features. This integration aims to mitigate data imbalance and identify feature subsets that maximize the F1- measure, thereby enhancing the dataset's overall quality and ensuring that feature interactions contributing to predictive performance are effectively captured. The second stage employs a multitasking feature selection framework based on the Grey Wolf Optimizer (GWO), which involves feature ranking and tiering to explore the dependencies between features and the target label. By categorizing features into different tiers and sampling them to generate tasks, this stage aims to maximize the dependency between features and the label. Additionally, the proposed pioneering knowledge transfer strategy facilitates effective knowledge sharing among tasks during the initialization phase, further refining the feature selection process. Tested across multiple imbalanced datasets, the algorithm demonstrates significant improvements in accuracy, F1 score, recall, and precision. This approach enhances model performance on imbalanced datasets and contributes valuable insights and methodologies.
Keywords:
Multitasking optimization
Feature selection
Imbalanced data
Knowledge transfer
Particle swarm optimization
Grey wolf optimization
Journal
IF:
8.5
Papers:
2.2K
Citations:
1.0W