Social Risk Early Warning of Environmental Damage of Large-Scale Construction Projects in China Based on Network Governance and LSTM Model

Table 1

Social risk early warning index system for environmental damage of large-scale construction projects.

Primary indicator	Secondary indicators	Definitions
Risk perception	Online search	Baidu Index reflects the exposure and user attention of different keywords in the past period and describes social hotspots, interests, and needs of all circles.
	Original Weibo	Original Sina Weibo related to environmental damage of large-scale construction project.
	Forwarding	The number of reposts on Weibo indicates that they agree with the opinions of original users.
	Comment	The number of comments on Weibo indicates that they agree with the opinions of original users.
Performance of risk information	Point of view	Refers to the user’s attitude towards the environmental damage of the project, such as approval, neutrality, and opposition. This indicator is mainly expressed by the ratio of negative expression to all Weibo.
	Information authenticity	The percentage of users who certify users through Sina. In generally, the authenticated user is responsible for the authenticity of his speech and will not distribute false information easily.
	Risk transmission	Refers to the user’s willingness to share information with other users. The higher the frequency of @ other people, the wider the scope of information dissemination.
Risk diffusion state	The gradient of online search	(Q2 − Q1)/(T1 − T2). The online search amount at time T1 is Q1, and T2 is Q2.
	The gradient of original Weibo	(Q2 − Q1)/(T1 − T2). The original Weibo amount at time T1 is Q1, and T2 is Q2.
	The gradient of forwarding	(Q2 − Q1)/(T1 − T2). The forwarding amount at time T1 is Q1, and T2 is Q2.
	The gradient of comment	(Q2 − Q1)/(T1 − T2). The comment amount at time T1 is Q1, and T2 is Q2.