Name
Title | ||
|---|---|---|
A Knowledge Enhanced Ensemble Learning Model for Mental Disorder Detection on Social Media. |
Abstract | ||
|---|---|---|
With the rapid development of social media in recent decades, a large amount of data posted by people with mental disorders has become available for researchers to use deep learning methods to detect potential mental disorders automatically. However, the current methods usually neglect to consider commonsense knowledge for detection tasks. Meanwhile, uneven data distribution is a great challenge for the current models. To address these problems, we propose a knowledge enhanced ensemble learning model for mental disorder detection on social media. First, we inject the knowledge triples into DailyPosts (Posts in one day posted online by a user) and then pass the reconstructed DailyPosts through a hierarchical model. The hierarchical model employs the BERT (Bidirectional Encoder Representations from Transformers) as the word embedder to integrate the BiGRU (Bidirectional Gated Recurrent Unit) network with an attention mechanism. Second, we combine the knowledge enhanced hierarchical model with the AdaBoost ensemble learning algorithm for the data imbalance. The proposed model was evaluated on two mental disorders (depression and anorexia) detection tasks. The model achieved competitive performance for depression and anorexia detection. The experimental results indicate that our model can obtain a robust performance in realistic scenarios. |
Year | DOI | Venue |
|---|---|---|
2020 | 10.1007/978-3-030-55393-7_17 | KSEM (2) |
DocType | Citations | PageRank |
Conference | 1 | 0.40 |
References | Authors | |
0 | 5 | |
Name | Order | Citations | PageRank |
|---|---|---|---|
| Guozheng Rao | 1 | 30 | 9.79 |
| Chengxia Peng | 2 | 1 | 0.40 |
| Li Zhang | 3 | 1 | 0.73 |
| Xin Wang | 4 | 2 | 1.82 |
| Zhiyong Feng | 5 | 794 | 167.21 |