游戏类淘宝直播带货违法监测预警数据

对采集的淘宝平台带货品类为游戏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：玩到停不下来、不限时、一夜暴富、学生、真金白银、免费兑换、18禁、充的越多越厉害、百万收入不是梦、畅玩等）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淘宝直播行为，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 live stream video content of Taobao platform influencers promoting game-related products, and processes and analyzes the spoken language content of the influencers. Based on the frequency and number of violations of pre-set sensitive violation keywords (such as 'can't stop playing', 'unlimited time', 'get rich overnight', 'students', 'real money', 'free redemption', '18+ restricted', 'the more you top up, the stronger you become', 'a million-dollar income is not a dream', 'unrestricted gameplay', etc.) by the influencers during live streams, warnings or corresponding treatments will be issued according to the trigger condition rules. This dataset provides data support for local Market Supervision Bureaus to regulate the live streaming behaviors of enterprises under their jurisdictions on Taobao.

The collected live stream videos are preprocessed through the following five steps:
1. Slice the original video files into segments with a maximum duration of 10 minutes;
2. Extract speech audio from the sliced videos via video-to-speech conversion;
3. Convert the extracted speech audio into text using automatic speech recognition (ASR) technology;
4. Extract text content from images captured from the original videos with Optical Character Recognition (OCR) technology;
5. Match the obtained text content against the pre-built violation warning keyword database.

Finally, a multi-criteria decision analysis (MCDA) model is applied to analyze and calculate the violation statements made by the anchors during live streams, so as to derive the violation warning score and the early warning judgment result. No early warning prompt is triggered when the violation warning score is ≤1, and a violation early warning prompt is activated when the score is >1.

搜集汇总

数据集介绍

特点

该数据集主要用于监测淘宝平台游戏类直播带货中的违法行为，包含1090条数据，每季度更新一次。数据集通过分析直播内容中的违规敏感词，如'玩到停不下来'、'一夜暴富'等，结合算法规则计算违法预警值，为市场监督局提供数据支持。