Decision Tree based System on Chip for Forest Fires Prediction

We expose in this work, the decision tree based intellectual property (IP) core development for forest fires prediction. We introduce its integration into the MicroBlaze based SoC architecture that constitutes the processing part of the sensor node. The aim is to speed up the predicting process by giving the decision locally at sensor node level. The decision tree based predictive model is simulated and trained using MATLAB tool for the tree generation; prior to its hardware development using the high level synthesis approach. The performance of the decision tree classifier in terms of accuracy and recall are about 75% and 0.88, respectively. The hardware implementation results of the decision tree based forest fires prediction system on chip show that the developed IP core requires few resources.