基于多维度数据的帖文推广检测方法、装置及电子设备
By analyzing data from multiple dimensions, we can obtain information on post content, interaction types, trends, and authors, thus solving the accuracy problem of promotional post detection and achieving more efficient promotional detection.
CN117435812BActive Publication Date: 2026-07-17ONE NETWORK INTEROPERABILITY (BEIJING) TECH CO LTD
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- ONE NETWORK INTEROPERABILITY (BEIJING) TECH CO LTD
- Filing Date
- 2023-11-10
- Publication Date
- 2026-07-17
AI Technical Summary
Technical Problem
In existing technologies, the detection of promotional posts suffers from problems such as difficulty in determining payment status, misjudgment due to limited data dimensions, and difficulty in detecting adjustments to promoter strategies.
Method used
Multi-dimensional data analysis is used to obtain information such as post content, interaction type, interaction trend, author profile, and other posts by the author. Feature extraction and weight calculation are used to determine whether a post contains promotional attributes.
Benefits of technology
It improves the accuracy of post promotion detection, reduces false negatives and false positives, adapts to the data characteristics of different social media platforms, and expands the scope of detection support.
✦ Generated by Eureka AI based on patent content.
Smart Images
Figure CN117435812B_ABST
Abstract
本发明实施例中提供了一种基于多维度数据的帖文推广检测方法、装置及电子设备，属于数据处理技术领域，该方法包括：获取与待检测帖文相关的帖文内容U1、互动类型U2、互动趋势U3、作者画像U4以及作者其他帖文U5，形成待检测帖文的检测数据集合U={U1,…U5}；利用特征提取公式g，对检测数据集合U进行特征计算，得到检测数据集合对应的特征值向量K=[K1,…K5]；获取待检测帖文的发布平台，以便于基于所述发布平台的历史数据，确定特征值集合K所对应的权重向量；基于所述特征值向量和所述权重向量，确定所述待检测帖文是否包含推广属性。采用本方案，能够基于多个维度对推广帖文进行检测，提高检测的准确性。
Need to check novelty before this filing date? Find Prior Art