Asset Details
Do you wish to reserve the book?
Mining Suicidal Ideation in Chinese Social Media: A Dual-Channel Deep Learning Model with Information Gain Optimization
Hey, we have placed the reservation for you!
By the way, why not check out events that you can attend while you pick your title.
Oops! Something went wrong.
Looks like we were not able to place the reservation. Kindly try again later.
Are you sure you want to remove the book from the shelf?
Mining Suicidal Ideation in Chinese Social Media: A Dual-Channel Deep Learning Model with Information Gain Optimization
Do you wish to request the book?
Mining Suicidal Ideation in Chinese Social Media: A Dual-Channel Deep Learning Model with Information Gain Optimization
Please be aware that the book you have requested cannot be checked out. If you would like to checkout this book, you can reserve another copy
We have requested the book for you!
Your request is successful and it will be processed during the Library working hours. Please check the status of your request in My Requests.
Oops! Something went wrong.
Looks like we were not able to place your request. Kindly try again later.
Journal Article
Mining Suicidal Ideation in Chinese Social Media: A Dual-Channel Deep Learning Model with Information Gain Optimization
2025
Overview
The timely identification of suicidal ideation on social media is pivotal for global suicide prevention efforts. Addressing the challenges posed by the unstructured nature of social media data, we present a novel Chinese-based dual-channel model, DSI-BTCNN, which leverages deep learning to discern patterns indicative of suicidal ideation. Our model is designed to process Chinese data and capture the nuances of text locality, context, and logical structure through a fine-grained text enhancement approach. It features a complex parallel architecture with multiple convolution kernels, operating on two distinct task channels to mine relevant features. We propose an information gain-based IDFN fusion mechanism. This approach efficiently allocates computational resources to the key features associated with suicide by assessing the change in entropy before and after feature partitioning. Evaluations on a customized dataset reveal that our method achieves an accuracy of 89.64%, a precision of 92.84%, an F1-score of 89.24%, and an AUC of 96.50%, surpassing TextCNN and BiLSTM models by an average of 4.66%, 12.85%, 3.08%, and 1.66%, respectively. Notably, our proposed model has an entropy value of 81.75, which represents a 17.53% increase compared to the original DSI-BTCNN model, indicating a more robust detection capability. This enhanced detection capability is vital for real-time social media monitoring, offering a promising tool for early intervention and potentially life-saving support.
