A real-time smoke early warning method based on spark and improved random forest

Abstract

In order to control and prevent fires in advance and effectively reduce the adverse effects of fires, this paper proposes a smoke early warning method based on spark and an improved random forest model. On the basis of the existing Internet of Things acquisition equipment, the real-time reception of the collected data is realized through spark streaming, and the collected data is persisted, and the trained model is used to judge the smoke early warning. The model established by this method is an improved random forest implementation based on the dragonfly optimization algorithm based on the samples collected in various environments. In the data preprocessing, the problem of data imbalance was found in the data set, and the oversampling method was used to solve it. Then this paper analyzes the problems in cross validation after data oversampling and proposes KSMOTE algorithm to solve this problem, which effectively improves the classification ability of the model. The experimental results show that the system has good real-time performance and accuracy.