典型区域火灾风险数据的特征挖掘研究

Research on Characteristic Mining of Fire Risk Data in Typical Areas

随物联网与大数据的技术发展,我国城市火灾风险防治工作日益智能化。城市社会治理和日常消防工作中积累了大量多维度、动态的数据,为更有价值的数据挖掘提供了基础。故对政府公开的基础数据和消防部门业务平台的专业数据,采用数据挖掘方法进行深入分析研究。首先,描述性分析数据的统计分布特征;其次,采用主成分分析、决策树信息增益算法,研究火灾风险与环境因素数据的关系特征,构建降维分析模型;最后,深化分析模型输出的数据规律,研究数据集各属性间变化的特征趋势,为消防大数据建设和应用提出建议。

With the technological development of the IoT and big data,the prevention and control of urban fire risk in China becomes increasingly intelligent.A large amount of multi-dimensional and dynamic data has been accumulated in urban social governance and daily firefighting work,which provides the basis for more valuable data mining.In this manuscript,the data which obtained from fire department business platform,government official website etc.were comprehensively analyzed and studied using data mining method.Firstly,a descriptive statistical analysis of databases was made by the visual display.Secondly,PCA(principal component analysis)and Decision Tree models were used to study the relationship between fire risk and environmental factors data,and the dimension reduction analysis model was built.Finally,by analyzing the data rule of the model output,the characteristic trend among the attributes of the data set was studied,and the valued suggestions were introduced for the construction and application of fire protection big data.