Scholarlink

The main aim is to compare Conventional and DL methods in mental health disorder prediction. This research includes the analysis of two different models of deep learning like CNN and RNN and machine learning models like Support Vector Machine, Random Forest, ADA Boosting, and Naive Bayes algorithm. Models are tested on three different datasets: Student_Behavior (records 40,960, attributes 25), Student mental health (records 101, attributes 11), Sleep_health_and_lifestyle_dataset (records 374, attributes 13) have been applied to the task of prediction of mental health illness. They were tested for classification performance measures like accuracy, precision, F1-score, recall. Experimental results show comparison between models. Here, ML models improve classification accuracy as compared to deep learning models. Naive Bayes achieves highest average accuracy among all the classifiers, i.e., 85.60%.

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 Exploration of Conventional Machine Learning and Deep Learning Classifier for Prediction of Mental Health Disorder

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Priyanka Mangal,
  • Harish Patidar

摘要

The main aim is to compare Conventional and DL methods in mental health disorder prediction. This research includes the analysis of two different models of deep learning like CNN and RNN and machine learning models like Support Vector Machine, Random Forest, ADA Boosting, and Naive Bayes algorithm. Models are tested on three different datasets: Student_Behavior (records 40,960, attributes 25), Student mental health (records 101, attributes 11), Sleep_health_and_lifestyle_dataset (records 374, attributes 13) have been applied to the task of prediction of mental health illness. They were tested for classification performance measures like accuracy, precision, F1-score, recall. Experimental results show comparison between models. Here, ML models improve classification accuracy as compared to deep learning models. Naive Bayes achieves highest average accuracy among all the classifiers, i.e., 85.60%.