Abstract:
To analyze the public opinion related to the employment situation, a combined approach is proposed to study the valuable ideas from social media. Firstly, the popularity of public opinion was analyzed according to the time series from a statistical point of view. Secondly, the feature extraction was carried out on the public opinion information, and the thematic analysis of the employment environment was carried out based on the Latent Dirichlet Allocation model. Thirdly, the Bert model was used to analyze the sentiment classification and trend of the employment-related public opinion data. Finally, the employment public opinion texts in different regions were studied based on the spatial sequence popularity analysis, keyword difference analysis. A case study in China is conducted to verify the effectiveness of proposed combined approach. Results shown that the popularity of employment public opinion reached the highest level in March 2022. Public opinions towards employment situation are negative. There is a specific relationship between the popularity of employment public opinion in different provinces.
摘要:
分析与就业形势有关的舆论,提出了一种组合方法来研究社交媒体的有价值的想法。首先,从统计的角度根据时间序列分析了舆论的流行程度。其次,对舆情信息进行特征提取,并基于潜在狄利克雷分配模型对就业环境进行了专题分析。第三,采用Bert模型对就业相关舆情数据进行情绪分类和趋势分析。最后,基于空间序列流行度分析,研究了不同地区的就业舆情文本,关键词差异分析。在中国进行了案例研究,以验证所提出的组合方法的有效性。结果显示,2022年3月,就业舆论的热度达到最高水平。公众对就业形势的看法是负面的。不同省份就业舆论的流行程度之间存在特定的关系。