Prediction and Analysis of Forest Fire using Meteorological Parameters

Prediction, prevention, and control of forest fires are becoming increasingly crucial for dealing with wildfires on all scales. Developing more effective fire detection systems can aid in the control of environmental threats. For forest fire prediction, the presented study used a kernel ridge regression approach with fuzzy logic (random search) based on meteorological measurements (temperature, wind direction, relative humidity, etc), soil moisture indicators, and geographical data. For the scoring matrix, mean absolute error (MAE) and root mean square error (RMSE) are used, and cross-validation 10 folds is performed to tune the hyper parameters. The result obtained by the presented model has a mean absolute error (MAE) of 9.89 and root mean square error (RMSE) of 23.38, which is minimum as compared to the prior works which means minimum the value of mean absolute error (MAE) and root mean square error (RMSE) more accurate will be the model with high predictive power. This paper analyses kernel ridge regression's significant impact on forecasting forest fire compared with different models with advancements made in technique with the help of substantial parameters. This technique (kernel ridge regression) can aid in the control of environmental threats by effective fire detection systems