Scholarlink

This exploration explores the development of a machine literacy-ground system for detecting anxiety through textbook and speech analysis. Anxiety diseases are one of the most current internal health issues, but early discovery remains grueling. Our approach leverages natural language processing (NLP) and audio signal processing to dissect textual and oral cues, aiming to give an effective and non-invasive tool for relating anxiety. The textbook analysis element focuses on verbal patterns, sentiment, and semantic features, while the speech analysis examines prosodic features similar to pitch, tone, and speech rate, which are frequently affected by anxiety. We trained and tested multiple machine literacy models, including support vector machines (SVM), arbitrary timbers, and deep literacy infrastructures, on datasets containing both textual and speech data labeled for anxiety. Our system achieved high delicacy in detecting anxiety-related patterns with an accuracy of 96%, demonstrating the eventuality of integration into clinical and tone-monitoring operations. This work highlights the promising part of AI in internal health diagnostics and opens avenues for real-time, automated anxiety discovery tools.

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

Dual-Stream Anxiety Detection: LSTM-Based Speech Analysis and CNN-Driven Sentiment Analysis

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Adrija Basu,
  • Akanksha Mishra,
  • Aditya Roy,
  • Aparijit Chakraborty,
  • R. Hariharan

摘要

This exploration explores the development of a machine literacy-ground system for detecting anxiety through textbook and speech analysis. Anxiety diseases are one of the most current internal health issues, but early discovery remains grueling. Our approach leverages natural language processing (NLP) and audio signal processing to dissect textual and oral cues, aiming to give an effective and non-invasive tool for relating anxiety. The textbook analysis element focuses on verbal patterns, sentiment, and semantic features, while the speech analysis examines prosodic features similar to pitch, tone, and speech rate, which are frequently affected by anxiety. We trained and tested multiple machine literacy models, including support vector machines (SVM), arbitrary timbers, and deep literacy infrastructures, on datasets containing both textual and speech data labeled for anxiety. Our system achieved high delicacy in detecting anxiety-related patterns with an accuracy of 96%, demonstrating the eventuality of integration into clinical and tone-monitoring operations. This work highlights the promising part of AI in internal health diagnostics and opens avenues for real-time, automated anxiety discovery tools.