社交媒体抑郁检测及情感分析数据集

为了提升抑郁检测的准确性与效果，本项目从多个公开的社交媒体平台中收集并筛选了相关的抑郁检测和情感分析数据集，涵盖中英文两种语言。这些数据集包括来自微博、推特、Reddit等国内外知名社交媒体平台的内容，旨在通过分析社交媒体上的用户言论来识别抑郁症状。抑郁检测部分的数据集包含三种标签：未抑郁、轻度抑郁和重度抑郁，帮助区分不同程度的抑郁状况；而情感分析数据集则包含积极与消极两个标签，用于识别用户情绪的基本倾向。
为了进一步优化抑郁检测效果，项目在抑郁检测的基础上，额外引入了情感分析数据集，以实现知识共享的辅助作用。通过对情感分析的结果进行结合，能够更加精准地识别潜在的抑郁症状。例如，通过情感分析来区分消极情绪的用户，进一步帮助判断其是否可能存在抑郁倾向。
情感分析数据集特别选择了中文情感分析数据集，该数据集主要来源于微博热搜，收集了大量的热点讨论和用户评论。这些数据集的引入，不仅增强了情感分析模型的适用性，还能提供更多的上下文信息，提升情感和抑郁检测模型的准确度。通过这些来自不同社交平台、多语言、多标签的数据，项目能够为抑郁症的早期检测提供强有力的数据支持，助力精准诊断和个性化干预。

To improve the accuracy and performance of depression detection, this project collected and curated relevant depression detection and sentiment analysis datasets from multiple public social media platforms, spanning both Chinese and English languages. These datasets contain content from well-known domestic and international social media platforms such as Weibo, Twitter, and Reddit, with the goal of identifying depressive symptoms by analyzing user-generated content on social media. The depression detection datasets include three labels: non-depressed, mild depression, and severe depression, which enable the distinction of different degrees of depressive conditions; while the sentiment analysis datasets feature two labels: positive and negative, used to identify the basic emotional tendency of users.

To further optimize the performance of depression detection, this project additionally introduced sentiment analysis datasets based on the depression detection datasets to serve as an auxiliary tool for knowledge sharing. By integrating the results of sentiment analysis, potential depressive symptoms can be identified with higher accuracy. For instance, distinguishing users with negative emotions via sentiment analysis can further assist in determining whether they may have depressive tendencies.

Specifically, one of the selected sentiment analysis datasets is a Chinese dataset mainly sourced from Weibo Hot Search, which collects a large number of hot topic discussions and user comments. The introduction of these datasets not only enhances the applicability of sentiment analysis models but also provides additional contextual information, improving the accuracy of both sentiment and depression detection models. Leveraging these multi-platform, multi-lingual, and multi-label datasets, this project can provide robust data support for the early detection of depression, facilitating accurate diagnosis and personalized intervention.

搜集汇总

背景与挑战

背景概述

该数据集是一个用于社交媒体抑郁检测及情感分析的多语言数据集，包含来自微博、推特等平台的用户言论，标注了抑郁程度和情感倾向。数据集由中国科学技术大学创建，旨在通过结合抑郁检测和情感分析提升早期抑郁症状识别的准确性。