Scholarlink

Authentic text feedback is crucial for informed decisions, while fake reviews distort customer perceptions, undermining the benefits of digital feedback. Therefore, it is crucial to detect and address fraudulent reviews. Despite limited success, digital detection and tracking in this challenging task need further exploration. Our experimental work introduces an enhanced machine learning model, utilizing the ABC algorithm to optimize accuracy through ensemble learning on public datasets via ML classifiers. Computer-based ML classifiers significantly outperform individual reviewers in identifying fraudulent reviews. Our findings strongly indicate that while human identification of fake reviews poses challenges, machine-based approaches excel in detecting deceptive content. These classifiers showed high accuracy rates between 80–95%, outperforming similar techniques. The findings showed lower MAE (0.015–0.4) and RMSE (0.08–0.9) values compared to previous studies. Our framework successfully identifies deceptive reviews via NLP. Moreover, we are comparing methods to find better spam detection approaches.

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

A Spam Detecting Model Based on Basic ML Classifiers: Comparative Analysis via ABC Algorithm and Result Generation

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Irtiqa Amin,
  • Harpreet Kaur,
  • Quraazah Akeemu Amin

摘要

Authentic text feedback is crucial for informed decisions, while fake reviews distort customer perceptions, undermining the benefits of digital feedback. Therefore, it is crucial to detect and address fraudulent reviews. Despite limited success, digital detection and tracking in this challenging task need further exploration. Our experimental work introduces an enhanced machine learning model, utilizing the ABC algorithm to optimize accuracy through ensemble learning on public datasets via ML classifiers. Computer-based ML classifiers significantly outperform individual reviewers in identifying fraudulent reviews. Our findings strongly indicate that while human identification of fake reviews poses challenges, machine-based approaches excel in detecting deceptive content. These classifiers showed high accuracy rates between 80–95%, outperforming similar techniques. The findings showed lower MAE (0.015–0.4) and RMSE (0.08–0.9) values compared to previous studies. Our framework successfully identifies deceptive reviews via NLP. Moreover, we are comparing methods to find better spam detection approaches.