青海典型可持续社会安全事件舆情数据

采取网络舆情分析与实地调研相结合的方式，识别青藏高原地区可持续发展过程中的社会安全风险。社会安全事件发布者的社会影响力和关注度也是影响舆情数量与质量的重要方面。选取央视新闻、新华每日电讯、红星新闻、新闻快报、人民数据、一锤视频、青海网络广播电视台、西宁晚报、青海法制报、政事儿、央广网、青海消防、封面新闻等14家媒体作为社会安全事件短视频发布者，它们都是拥有一定规模数量的粉丝和网络影响的官方权威媒体。
社会安全事件数据集来源于网络爬虫。为了提高舆情数据分析的工作效率，本文使用自然语言处理方法进行可持续社会安全事件舆情数据的批量处理和分析。该方法是当前计算机科学与人工智能分析结合的一个重要领域，主要应用于机器自动翻译、舆情动态监测、网民观点提取、词句分门别类、文本语义分析比对等方面的科学研究。情感属性分析作为自然语言处理的主要功能，它能够很好地对情感色彩比较浓厚的主观文本进行分析、处理、归纳和推理，准确识别出不同社会群体对某一社会事件的情感倾向（积极属性、消极属性及中性属性），由此可以计算出各种情感倾向所占据的比例。

This study adopts a combination of online public opinion analysis and field research to identify social security risks during the sustainable development of the Qinghai-Tibet Plateau region. The social influence and attention of the publishers of social security incidents are also important factors affecting the quantity and quality of public opinions. Fourteen media outlets including "CCTV News", "Xinhua Daily Telegraph", "Red Star News", "News Express", "People's Data", "Yichui Video", "Qinghai Network Radio and Television Station", "Xining Evening News", "Qinghai Legal Daily", "Zhengshiye", "China National Radio Network", "Qinghai Fire Rescue", and "Cover News" were selected as publishers of short videos about social security incidents. All these are official and authoritative media with a certain scale of followers and online influence. The social security incident dataset is sourced from web crawlers. To improve the efficiency of public opinion data analysis, this paper uses Natural Language Processing (NLP) methods to perform batch processing and analysis of public opinion data related to sustainable social security incidents. This method is an important research field that combines computer science and artificial intelligence analysis, and is mainly applied in scientific research such as machine translation, dynamic public opinion monitoring, extraction of netizens' opinions, classification of words and phrases, text semantic analysis and comparison, etc. As a core function of NLP, sentiment attribute analysis can well analyze, process, summarize and reason about subjective texts with strong emotional colors, and accurately identify the emotional tendencies (positive, negative and neutral attributes) of different social groups towards a certain social event, so as to calculate the proportion occupied by various emotional tendencies.

搜集汇总

背景与挑战

背景概述

该数据集包含2018年至2023年青海省典型可持续社会安全事件的舆情数据，通过14家官方权威媒体的短视频和网络爬虫获取，并利用自然语言处理方法进行情感属性分析，用于识别社会安全风险和舆情动态监测。