【摘要】 针对常见文本中情绪倾向性不明显的问题，提出一种融合BERT模型与网络爬虫技术的应对方法。利用网络爬虫技术搜集社交网络的数据资料为情感分析提供实例和数据支撑，采用交错验证法将数据集分割为训练子集和预测子集。利用数据集进行BERT学习和文本分类，即运用深度学习和自然语言处理对文字资料实施归类。研究结果表明，采用BERT模型实行情感分析的技术在精确度与广泛适用性上展现出良好的性能，正确率可达97.86%。更多还原

【Abstract】 For the problem that the emotional tendency is not obvious in common texts, this paper proposes a coping method combining BERT model and web crawler technology. The use of web crawler technology to collect data from social networks provides examples and data support for emotion analysis. Interleaved validation was used to split the datasets into training and prediction subsets. Using data sets for BERT learning and text classification, that is, deep learning and natural language learning are used to classify text data. The study showed that the use of BERT model to implement emotion analysis showed good performance in accuracy and wide applicability, with an accuracy rate of 97.86%.更多还原