基于文本挖掘的公众对绿色建筑的情感分析

吴露洁

摘 要

由于环境可持续发展的需求，绿色建筑已经成为我国建设转型的方向，然而，目前我国的绿色建筑推广还停留在政府过度干预、市场参与度不足的老路上，尽管中国政府为推动绿色建筑的发展做出了很多努力，但其发展仍旧十分缓慢，绿色建筑规模化推广难以实现。现阶段，鲜少有相关的研究来调查公众对于绿色建筑的认知和态度。

为了探究中国公众对绿色建筑的关注程度、变化趋势、情感取向和关注焦点，本文收集并分析了绿色建筑相关的微博用户信息以及热门帖子和评论，主要采用网络爬虫技术，结合LDA主题建模和情感分析两种文本挖掘方法，探索公众对绿色建筑的负面情绪焦点，具体如下：

首先，利用jieba分词系统，对爬取的数据文本进行分词等预处理，一方面，对评论文本进行LDA主题建模，挖掘出不同类型用户的关注焦点。另一方面，运用词性模板的方法，提取特征词和情感词，通过人工标注，建立起绿色建筑领域的情感词典，并给出情感词典各要素的情感强度标注处理方法。论文构建的情感词典要素主要包括情感词、否定词、程度副词以及评论的搭配关系，在对这些元素进行情感强度标注结束后，对评论文本进行情感倾向计算。最后，将最终的评论短语以可视化词云的方式展示，为绿色建筑的政策制定及推广方向提供相关的建议。

研究结果显示，公众对绿色建筑的关注度并不高，微博上绿色建筑话题的发帖量和粉丝数量都不够多，在人们的认知中，绿色建筑主要表现为垂直绿化住宅、被动式房屋以及装配式建筑，在对“绿色”的理解上，人们的认知还存在一定的偏差，很多人认为绿色建筑就是绿色的建筑，因此对绿色植物带来的蚊虫产生了排斥心理，而在被动式和装配式建筑中，人们的负面情绪主要集中在技术和安全的问题上。

关键词：文本挖掘；绿色建筑；情感分析；新浪微博

Abstract

Because of the demand for the sustainable development of environment, green building has become the direction of construction in our country, however, at present, China's green building promotion with excessive government intervention, lack of market participation, on the path of although the Chinese government made a lot of effort to promote the development of green building, but its development is still very slow.At present, there are few relevant studies to investigate the public's perception and attitude towards green buildings.

In order to explore the Chinese public's attention in the green building, trends, emotional orientation and focus, this paper collects and analyzes the green building related weibo user information and popular posts and comments, mainly USES the web crawler technology, combined with the LDA theme modeling and emotional analysis of two methods of text mining, explore the public's negative emotional focus toward the green building, specific as follows:

First, jieba word segmentation system is used to preprocess the crawled data text into word segmentation. On the one hand, LDA topic modeling is carried out on the comment text to mine the focus of different types of users.On the other hand, the method of part of speech template is used to extract feature words and emotion words, and the emotion dictionary in the field of green building is established by manual labeling, and the method of labeling the emotion intensity of each element of the emotion dictionary is given.The emotional dictionary elements constructed in this paper mainly include the collocation relations of emotional words, negative words, adverbs of degree and comments. After marking the emotional intensity of these elements, the emotional tendency of the comment text is calculated.Finally, the final comment phrase is presented in the form of visual word cloud to provide relevant Suggestions for the policy development and promotion direction of green building.

Results show that public awareness is not high, for green building green building topic on weibo postings and fans quantity are not enough, in the people's cognitive, green construction mainly for vertical greening residential, passive houses and prefabricated construction, on the understanding of "green", people's cognitive also has certain deviation, many people believe that green buildings is green, so on green plants of mosquito resistance, in the passive and prefabricated buildings, the negative emotions of people focused on technology and safety problems.

Key words: text mining; green building; emotion analysis; sina weibo