Scholarlink

Performance efficiency is greatly affected by the choice of model, especially as data complexity increases. Since even minor increases in accuracy can result in better decision-making, where accurate classification is essential for a variety of real-world applications. This study evaluates multiple classification models, including deep Convolutional neural network (CNN), Multilabel classifier, CNN (VGG-Inception), Naïve Bayes, Decision tree, Support vector machine (SVM), and Optimized ant gradient-CNN (OAG-CNN), to compare the efficiency guided by essential metrics as accuracy, precision, sensitivity, F1-score, and computational time. The Apple leaf disease dataset (ALDD) underwent preprocessing approaches such as data augmentation and normalization to enhance training effectiveness. Experimental results indicate that deep learning (DL) models, particularly CNN-based architectures, outperform traditional machine learning (ML) approaches. Among the evaluated models, OAG-CNN obtained the highest accuracy of 98.5% while maintaining the lowest computation time of 2 s, making it a strong candidate for real-time classification tasks. These outcomes provide meaningful findings into enhancing the efficiency and accuracy of DL-based classification, helping to refine model selection and optimization for real-world applications.

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

Comparative Analysis of OAG-CNN, Deep CNN, VGG, and Other Models for Enhanced Performance Evaluation

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Arshleen Kaur

摘要

Performance efficiency is greatly affected by the choice of model, especially as data complexity increases. Since even minor increases in accuracy can result in better decision-making, where accurate classification is essential for a variety of real-world applications. This study evaluates multiple classification models, including deep Convolutional neural network (CNN), Multilabel classifier, CNN (VGG-Inception), Naïve Bayes, Decision tree, Support vector machine (SVM), and Optimized ant gradient-CNN (OAG-CNN), to compare the efficiency guided by essential metrics as accuracy, precision, sensitivity, F1-score, and computational time. The Apple leaf disease dataset (ALDD) underwent preprocessing approaches such as data augmentation and normalization to enhance training effectiveness. Experimental results indicate that deep learning (DL) models, particularly CNN-based architectures, outperform traditional machine learning (ML) approaches. Among the evaluated models, OAG-CNN obtained the highest accuracy of 98.5% while maintaining the lowest computation time of 2 s, making it a strong candidate for real-time classification tasks. These outcomes provide meaningful findings into enhancing the efficiency and accuracy of DL-based classification, helping to refine model selection and optimization for real-world applications.