| With the gradual improvement of my country’s urbanization construction,reducing the incidence of urban fires and reducing casualties and losses has always been one of the major goals of urban fire protection construction.Existing fire early warning systems or smart fire fighting systems can alert in time when a fire occurs,but there are still many shortcomings:(1)The system lacks the support of big data technology and does not have the ability to centralize data management.(2)In the face of massive fire data,there is a lack of effective data mining and analysis applications.(3)Fire early warning is realized by Io T equipment,lack of fire prediction application or poor prediction effect.(4)The applicability of the fire protection unit safety risk assessment method used in urban fire prevention is not strong and the assessment cost is high.This paper uses big data and artificial intelligence technology to research and apply urban fire management and fire prevention.This paper main work is as follows:1.In response to the current weak data centralized management capabilities of the fire protection system,this article uses Hadoop,Storm,HBase,Kafka and other big data technologies to establish a unified data center with data collection,data processing,data distribution and data storage functions.It can provide accurate and complete basic fire protection data for upper-level applications,and provide strong data support for statistical analysis and predictive analysis.2.In order to explore the potential value of historical fire data,this paper proposes a new linear regression model to complete the prediction and analysis of the number of fires,and the average prediction error is within 7%.At the same time,this paper also proposes an improved fire prediction method based on LSTM.Compared with the traditional prediction method based on LSTM,this method has smaller prediction errors and higher prediction accuracy.3.This paper proposes a broader fire safety risk assessment method for a unit based on analytic hierarchy process.This method abstracts many risk indicators of the unit,and calculates the fire safety risk assessment score of the unit by combining the scoring formula and weight of each type of index.Reduce the burden of calculating index weights and reduce the cost of fire safety risk assessment. |