基于多层模糊神经网络的消防预警系统设计

Design of Fire Warning System Based on Multilayer Fuzzy Neural Network

针对城市消防预警能力有待提升的问题,分析了城市消防远程监控系统的数据构成特征。阐述了以消防巡查射频打卡数据和温度数据为挖掘重点的数据预警逻辑架构。采用多项式和对数深度迭代回归模糊算法,将输出值导入到预警结果整理模块后,通过模糊矩阵法生成模糊预警的方法,设计不同时间值域的本地双列数据。同时,约束当前时间点,并使用外部全城其他节点数据形成的参照矩阵,构建仿真设计方案。通过真实数据的仿真测试验证,该系统在不同消防预警级别下的敏感度、特异度均满足要求。与可查参考文献中其他机器学习算法对比发现,该系统的火情预警系统最优值相比更优。该系统拥有可置信的统计学优势。

Aiming at the problem that the urban fire warning capability needs to be improved,the data composition characteristics of urban fire remote monitoring system are analyzed.The logical architecture of data warning focusing on the mining of fire inspection radio frequency punch card data and temperature data is described.Polynomial and logarithmic depth iterative regression fuzzy algorithms are used to design local two-column data with different time value domains by means of fuzzy matrix method for generating fuzzy warning after importing the output values into the warning result collation module.At the same time,the current time point is constrained,and the reference matrix formed by external city-wide data from other nodes is used to construct the simulation design scheme.Through the simulation test with real data,it is verified that the sensitivity and specificity of the system under different fire warning levels meet the requirements.Compared with other machine learning algorithms in the available references reveals,the fire warning system optimal values of the system are better.The system possesses a credible statistical advantage.