Scholarlink

<p>This paper introduces a novel classification model for lung cancer prediction through preprocessing, feature selection (FS) algorithms, and deep learning (DL) models. The public shared dataset from the Kaggle repository was utilized, which consists of 15 attributes associated with lung cancer. Data preprocessing tasks included addressing issues such as missing values, label encoding, and standardization. SMOTE was used to address class imbalance. Five FS algorithms in binary version: Particle Swarm Optimization (BPSO), Genetic Algorithm, Grey Wolf Optimizer, Whale Optimization Algorithm, and Ant Colony Optimization were utilized to evaluate performance. The BPSO achieved the best accuracy with 98.15% to select 11 optimal features of 15. These selected attributes were employed to assess suggested DL models, such as Multilayer Perceptron (MLP), CNN, Recurrent Neural Network, Long Short-Term Memory, and Gated Recurrent Unit. A novel stacked model integrating a CNN and an MLP with a random forest meta-classifier was developed, achieving exceptional performance with 99% AUC, 98.7% accuracy, precision, recall, and F1-score. The results show that the stacked model outperforms individual models in classifying lung cancer. This study emphasizes the efficacy of FS and stacking methodologies in improving the predictive performance of DL models.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Enhancing lung cancer disease classification based on BPSO and stacked MLP with CNN model

  • 期刊论文

  • 发表日期：

    2026年5月13日

查看全文
  • Ahmed M. Elshewey,
  • Ahmed M. Osman,
  • Mohamed S. Sawah,
  • Hazem M. El-Bakry,
  • Samah A. Z. Hassan

摘要

This paper introduces a novel classification model for lung cancer prediction through preprocessing, feature selection (FS) algorithms, and deep learning (DL) models. The public shared dataset from the Kaggle repository was utilized, which consists of 15 attributes associated with lung cancer. Data preprocessing tasks included addressing issues such as missing values, label encoding, and standardization. SMOTE was used to address class imbalance. Five FS algorithms in binary version: Particle Swarm Optimization (BPSO), Genetic Algorithm, Grey Wolf Optimizer, Whale Optimization Algorithm, and Ant Colony Optimization were utilized to evaluate performance. The BPSO achieved the best accuracy with 98.15% to select 11 optimal features of 15. These selected attributes were employed to assess suggested DL models, such as Multilayer Perceptron (MLP), CNN, Recurrent Neural Network, Long Short-Term Memory, and Gated Recurrent Unit. A novel stacked model integrating a CNN and an MLP with a random forest meta-classifier was developed, achieving exceptional performance with 99% AUC, 98.7% accuracy, precision, recall, and F1-score. The results show that the stacked model outperforms individual models in classifying lung cancer. This study emphasizes the efficacy of FS and stacking methodologies in improving the predictive performance of DL models.