Souha-BH/DetectingRiskyHealthBehaviorsInTikTokVideos

Detecting Risky Health Behaviors in TikTok Videos数据集是一个包含TikTok视频元数据和注释的集合，旨在促进开发能够检测和分类短视频中风险健康行为的机器学习模型。数据集包括视频的唯一标识符、创建时间、评论数、点赞数、播放次数、分享次数、视频标题、视频URL、数据集划分、风险行为标签以及视频名称等信息。数据集通过Apify TikTok Hashtag Scraper工具从TikTok收集，使用了特定的健康相关标签，并限制了每个标签的视频数量。注释过程要求注释者观看视频并判断是否展示或提及风险健康行为。

The Detecting Risky Health Behaviors in TikTok Videos dataset is a collection of TikTok video metadata and annotations aimed at facilitating the development of ML models that can detect and classify risky health behaviors depicted in short-form TikTok videos. The dataset includes columns such as ID, createTimeISO, commentCount, diggCount, playCount, shareCount, text, webVideoUrl, Split, Risky Behavior, and Video Name. The dataset was collected from TikTok on October 1st, 2024 using the Apify TikTok Hashtag Scraper tool, with specific hashtags like #cigarettes, #alcohol, #diet, #workout, #food, #smoking, and limited to 100 videos per hashtag. The annotation process requires annotators to watch no more than 30 seconds of each video to determine if it depicts or mentions risky health behaviors.