Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

CautionSuicide: A Deep Learning Based Approach for Detecting Suicidal Ideation in Real Time Chatbot Conversation

Nelly Elsayed, Zag ElSayed, Murat Ozer

TL;DR

The paper tackles real-time detection of suicidal ideation in digital chat content by combining a lightweight GRU-based NLP model with a chatbot framework for automated screening and escalation. It leverages a large Reddit-derived dataset (SuicideWatch/Depression vs non-suicide) and a 100k-word vocabulary to achieve a test accuracy of 94.33% and an F1 score of 0.9433, outperforming state-of-the-art baselines. The framework integrates with chat interfaces and authority-reporting mechanisms to support timely intervention while noting privacy and security limitations. Overall, the work offers a practical, scalable approach to embedding mental-health monitoring in conversational agents with potential impact on education, medicine, and enforcement.

Abstract

Suicide is recognized as one of the most serious concerns in the modern society. Suicide causes tragedy that affects countries, communities, and families. There are many factors that lead to suicidal ideations. Early detection of suicidal ideations can help to prevent suicide occurrence by providing the victim with the required professional support, especially when the victim does not recognize the danger of having suicidal ideations. As technology usage has increased, people share and express their ideations digitally via social media, chatbots, and other digital platforms. In this paper, we proposed a novel, simple deep learning-based model to detect suicidal ideations in digital content, mainly focusing on chatbots as the primary data source. In addition, we provide a framework that employs the proposed suicide detection integration with a chatbot-based support system.

CautionSuicide: A Deep Learning Based Approach for Detecting Suicidal Ideation in Real Time Chatbot Conversation

TL;DR

The paper tackles real-time detection of suicidal ideation in digital chat content by combining a lightweight GRU-based NLP model with a chatbot framework for automated screening and escalation. It leverages a large Reddit-derived dataset (SuicideWatch/Depression vs non-suicide) and a 100k-word vocabulary to achieve a test accuracy of 94.33% and an F1 score of 0.9433, outperforming state-of-the-art baselines. The framework integrates with chat interfaces and authority-reporting mechanisms to support timely intervention while noting privacy and security limitations. Overall, the work offers a practical, scalable approach to embedding mental-health monitoring in conversational agents with potential impact on education, medicine, and enforcement.

Abstract

Suicide is recognized as one of the most serious concerns in the modern society. Suicide causes tragedy that affects countries, communities, and families. There are many factors that lead to suicidal ideations. Early detection of suicidal ideations can help to prevent suicide occurrence by providing the victim with the required professional support, especially when the victim does not recognize the danger of having suicidal ideations. As technology usage has increased, people share and express their ideations digitally via social media, chatbots, and other digital platforms. In this paper, we proposed a novel, simple deep learning-based model to detect suicidal ideations in digital content, mainly focusing on chatbots as the primary data source. In addition, we provide a framework that employs the proposed suicide detection integration with a chatbot-based support system.
Paper Structure (8 sections, 6 figures, 4 tables)

This paper contains 8 sections, 6 figures, 4 tables.

Table of Contents

  1. Introduction
  2. Suicide Detection in Chatbot Conversation Framework
  3. Suicide ideation Detection Model
  4. Experiment and Results
  5. Dataset
  6. Experimental Setup
  7. Limitations and Open Questions
  8. Conclusion

Figures (6)

  • Figure 1: The number of committed suicides each month from 2018 to 2023 according to the Center for Disease Contol and Prevention (CDC) data published in cdc_statistics2
  • Figure 2: The proposed framework for suicidal ideation detection in a chatbot conversationn.
  • Figure 3: The proposed model for the suicidal ideation detection diagram.
  • Figure 4: The training versus validation accuracies of the proposed suicide ideations detection model.
  • Figure 5: The training versus validation losses of the proposed suicide ideations detection model.
  • ...and 1 more figures