| 摘 要: 为了研究房地产文本评论情感对房地产价格波动的影响及其造成的市场现象，提出了基于CNN(卷积神经网络)和BiLSTM(双向长短期记忆网络)混合方法的房地产网络舆情分析模型。构建地域专属情感词典，通过CNN和BiLSTM获得文本情感极性，利用量化情感指数计算得到情感得分，将其与百度指数拟合计算互信息数。结果表明：经过量化的房地产文本情感值为67.84%与百度指数的相关系数值大致相同，二者走势基本吻合，利用该方法可实时测度房地产评论舆情走势，为数字经济背景下房地产企业决策者提供智力支持。 |
| 关键词: 房地产；网络舆情；情感指数；最大互信息 |
|
中图分类号: TP183
文献标识码: A
|
|
| Public Opinion Analysis of Real Estate Network based on Hybrid Deep Learning Method |
|
LI Shuaiwen, LIU Ji
|
(School of Statistics and Data Science, Xinjiang University of Finance and Economics, Urumchi 830012, China )
1691133906@qq.com; Liuji5000@126.com
|
| Abstract: In order to study the influence of real estate text comment sentiment on real estate price fluctuation and its market phenomenon, this paper proposes a real estate network public opinion analysis model based on the hybrid method of CNN (Convolutional Neural Network) and BiLSTM (Bi-directional Long Short-Term Memory). A region-specific sentiment dictionary is constructed and the text sentiment polarity is obtained through CNN and BiLSTM. The sentiment score is calculated by using the quantitative sentiment index, and the number of mutual information is calculated by fitting it to the Baidu index. The results show that the quantified emotional value of real estate text is 67.84%, which is roughly the same as the correlation coefficient value of Baidu Index. The trend of the two is basically consistent. This method can be used to measure the trend of public opinion of real estate reviews in real time, and provide intellectual support for decision-makers of real estate enterprises in the context of digital economy. |
| Keywords: real estate; network public opinion; sentiment index; maximum mutual information |
