基于微博情感序列的情感异常检测方法和系统
By constructing a three-dimensional sentiment vector matrix based on Weibo sentiment sequences and employing a multi-model training strategy, the problem of sentiment anomaly detection that cannot be universally applied to all individual users in existing technologies has been solved, thus achieving accurate detection of user sentiment.
CN117633629BActive Publication Date: 2026-07-17HEFEI UNIV OF TECH
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- HEFEI UNIV OF TECH
- Filing Date
- 2023-12-12
- Publication Date
- 2026-07-17
AI Technical Summary
Technical Problem
Existing technologies cannot be universally applied to the detection of emotional abnormalities in all individual users, resulting in poor detection results.
Method used
By collecting historical Weibo text sets with psychological risk level labels, and using GCN, TRANSFORMER, and CBOW models to extract sentiment vectors, a three-dimensional sentiment vector matrix is constructed. Then, an sentiment anomaly detection model is trained by combining LSTM neural networks and attention mechanisms to achieve accurate detection of user sentiment.
Benefits of technology
It is universally applicable to all individual users, accurately detecting emotional abnormalities in each user, thus improving the accuracy and universality of detection.
✦ Generated by Eureka AI based on patent content.
Smart Images
Figure CN117633629B_ABST
Abstract
本发明提供一种基于微博情感序列的情感异常检测方法、系统、存储介质和电子设备，涉及情感异常检测技术领域。本发明，通过有效利用用户的历史微博文本数据，综合从每一条历史微博文本中提取出简单情感向量、复杂情感向量和情感词情感向量，以获取二维情感向量矩阵；再以时间顺序将二维情感向量矩阵组合形成以三维情感向量矩阵为形式的情感序列；并在情感序列上构建并训练用户情感异常检测模型，从而能够普遍适用于所有用户个体，并精准地检测出每一个用户情感异常的情感异常检测方法。此外，还能够提高风险处理能力，在心理健康领域上具有较大优势；以及能够有效降低了情感表示的模糊度，提高检测精度。
Need to check novelty before this filing date? Find Prior Art