美容仪器类拼多多直播带货违法监测预警数据
收藏资源简介：
对采集的拼多多平台带货品类为美容仪类的达人直播视频内容进行转译分析，对达人口播语言内容进行处理、分析，根据达人在直播过程中对预先设置的违规敏感词（比如：全方位、超强、立刻见效、永久去除、院线级别、安全、无痛、不留疤痕、没有副作用等）违反的次数和频率，依据触发条件规则提出警告或处理。为服务辖区市场监督局管理区域内规范企业拼多多直播行为，提供数据支持。将采集完成的直播视频进行进行预处理，第一步：基于原始视频文件，以最大10分钟单位对原始视频进行切片。第二步：对于已完成的切片视频，进行视频内容转语音操作。第三步：对于已完成视频转语音操作的切片，进行语音转文本操作。第四步：使用OCR技术对原始视频中抓取的图片进行文字提取操作。第五步：将所得到的文字内容与违法预警关键词库进行匹配。最终运用多标准决策分析模型，对主播在直播过程中出现的违规语句进行分析计算，得出违法预警值和是否预警判断。违法预警值 ≤1 时，不触发预警提示，违法预警值 ＞1 时触发违法预警提示。
This dataset conducts transcription and analysis on influencer live stream videos for beauty equipment products collected from the Pinduoduo e-commerce platform. The spoken content of the influencers will be processed and analyzed, and warnings or corresponding treatments will be issued based on the frequency and count of violations against pre-set sensitive prohibited words during live broadcasts, including but not limited to: comprehensive, ultra-powerful, immediate results, permanent removal, professional salon-grade, safe, painless, no scarring, no side effects, etc. This work provides data support for local Market Supervision Bureaus to standardize the live streaming behaviors of enterprises within their jurisdiction on the Pinduoduo platform. The collected live stream videos undergo the following preprocessing steps: 1. Slice the original video files into segments with a maximum duration of 10 minutes per segment. 2. Perform video-to-speech conversion on the sliced videos to extract audio content. 3. Conduct speech-to-text conversion on the extracted audio files to generate transcribed text. 4. Extract text from images captured from the original videos using Optical Character Recognition (OCR) technology. 5. Match the collected text content against the pre-built prohibited keyword database for violation warnings. Finally, the Multi-criteria Decision Analysis (MCDA) model is applied to analyze and calculate the violative statements made by live stream hosts during their broadcasts, so as to derive the violation warning score and the corresponding warning judgment. Specifically, no warning alert will be triggered when the violation warning score is ≤ 1, while a violation warning alert will be triggered when the score is > 1.
