A coal mine gas concentration prediction method based on Particle Swarm Optimization Support Vector Regression

Gas concentration prediction plays an important role in the safety production of coal mine. Acceptable result has been achieved using Support Vector Regression to predict gas concentration; however, the efficiency of the algorithm is declined due to the parameter choice. This paper presents a coal mine gas concentration prediction method based on Particle Swarm Optimization Support Vector Regression which using Particle Swarm Optimization algorithm to optimize the parameters used in Support Vector Regression and improve the model´s accuracy and generalization ability. Experiments on the real coal mine gas concentration show that the proposed method is more accuracy than that without using Particle Swarm Optimization.