珍珠类阿里巴巴直播带货违法监测预警数据
收藏资源简介：
对采集的阿里巴巴平台带货品类为珍珠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：光滑细腻，天然眩光，孤品，s级，纯天然，925银防过敏，近珠光等）违反的次数和频率，依据触发条件规则提出警告或处理。为诸暨市市场监督局管理区域内规范企业阿里巴巴直播行为，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 dataset is developed based on collected live stream video content of influencers selling pearl products on the Alibaba Platform, with transcription and analysis conducted on the content. The oral speech content of the live streamers is processed and analyzed, and warnings or corresponding treatments are proposed in accordance with triggering rules, based on the times and frequency of violations of pre-set sensitive violation keywords (such as: "smooth and delicate", "natural glare", "one-of-a-kind item", "S-grade", "100% natural", "925 sterling silver anti-allergy", "near-perfect luster", etc.). This dataset provides data support for the Market Supervision and Administration Bureau of Zhuji City to regulate the live streaming behaviors of enterprises on the Alibaba Platform within its jurisdiction. The collected live stream videos undergo the following preprocessing procedures: Step 1: Slice the original video files into segments with a maximum duration of 10 minutes each. Step 2: Convert the sliced video content into audio files. Step 3: Convert the audio content into text transcripts. Step 4: Extract text from images captured in the original videos using OCR technology. Step 5: Match the obtained text content against the pre-built violation early-warning keyword database. Finally, a multi-criteria decision analysis (MCDA) model is utilized to analyze and calculate the violation statements made by the streamer during the live stream, so as to derive the final violation early-warning value and make a judgment on whether to trigger a warning prompt. The final violation early-warning value is calculated via the following formula: Violation Early-warning Value = (Number of hit single violation warning keywords * 0.25) + (Number of hit combined violation warning keywords * 0.3) + (Number of hit combined warning keywords from image recognition * 0.35) + (Number of historical violation records of the live stream room in the past month * 0.1) The warning triggering rule is as follows: No warning prompt will be triggered when the Violation Early-warning Value ≤ 1, and a violation warning prompt will be triggered when the Violation Early-warning Value > 1.
