Analysis of Fire Rescue Problems Based on Prediction and Fitting Algorithm

This paper studies the problem of fire rescue. According to the population, area, police situation and other data in different regions of the area, the number of police calls is predicted. Firstly, according to the data from 2016 to 2019, draw the image of the number of alarms changing with months; Then, polynomial model fitting and local weighted regression are used to predict the data in 2020 and compared with the real value; The square loss function is used to evaluate the accuracy of the model, and then the sensitivity analysis is carried out to evaluate the stability of the model; Finally, the model is used to predict the number of fire rescue calls in each month of 2021.