宠物食品类拼多多直播带货违法监测预警数据

对采集的拼多多平台带货品类为宠物食品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：绝对安全、预防、腹泻、皮肤病、寄生虫、增强免疫力、无添加、纯天然、治疗、无效退款、没有副作用、肯定有用等）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=（违法预警单关键词命中次数*0.25）+（违法预警组合关键词命中次数* 0.3）+（图片识别命中预警组合关键词个数*0.35）+（直播间近一个月历史违规记录数*0.1）
通过公式计算出最终违法预警值，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 work conducts transcription and analysis on collected influencer live stream videos of pet food products on the Pinduoduo platform. The spoken language content of the influencers is processed and analyzed. Warnings or penalties are issued based on the frequency and number of violations of pre-set sensitive prohibited keywords (such as absolutely safe, prevent, diarrhea, skin diseases, parasites, enhance immunity, no additives, purely natural, treatment, money-back guarantee, no side effects, definitely effective, etc.) during the live broadcast, in accordance with the trigger condition rules. This work provides data support for market supervision bureaus in their jurisdiction to standardize the live streaming behaviors of enterprises on the Pinduoduo platform.

The collected live stream videos undergo the following preprocessing steps:
1. Segment the original video files into clips with a maximum duration of 10 minutes.
2. Extract speech from the segmented video clips.
3. Perform speech-to-text conversion on the speech-extracted clips.
4. Use Optical Character Recognition (OCR) technology to extract text from images captured from the original videos.
5. Match the obtained text content against the prohibited early-warning keyword database.

Finally, a multi-criteria decision analysis (MCDA) model is employed to analyze and calculate the prohibited statements made by the streamer during the live broadcast, so as to derive the early-warning violation score and make a warning determination.
The early-warning violation score is calculated as follows:
Early-warning violation score = (Single keyword hit count of prohibited early-warning terms × 0.25) + (Combined keyword hit count of prohibited early-warning terms × 0.3) + (Number of combined early-warning keyword hits from image recognition × 0.35) + (Number of historical violation records of the live stream room in the past month × 0.1)
No warning prompt is triggered when the score ≤ 1, while a warning prompt is activated when the score > 1.

搜集汇总

数据集介绍

特点

该数据集用于监测拼多多平台上宠物食品类直播带货中的违法行为，通过分析直播内容中的违规关键词，计算违法预警值并判断是否触发预警。数据集规模为1436条，每季度更新，服务于市场监督局的监管工作。