study on a fire detection system based on support vector machine

It is very important to research the prediction of fire, which is significant to the people and nation. The traditional fire detection system based on neural network has the disadvantages of over learning, trapped in local minimum, etc. This paper proposes a new fire detection system based on support vector machine (SVM). Gas sensors, smoke sensor and temperature sensor are composed to be a sensor array. The fire detection model is established, including sample selection, prediction model training prediction, output modules, etc. The SVM transform the complicated nonlinear problem into the linear problem in the high dimensional plane. The experimental results show that fire detection system based on support vector machine had high recognition rate and reliability, it overcomes the disadvantages of traditional methods.