A Support Vector Machine model on correlation between the heterogeneous ignition temperature of coal char particles and coal proximate analysis

A prediction model of heterogeneous ignition temperature of coal char particles was built based on Support Vector Machine (SVM), in which there were four input vectors, which were moisture content, ash content, volatile content and fixed carbon of coal. A new optimization approach based on microscope principle was developed when identifying the optimal parameter pair of regularization parameter ¿ and kernel parameter ¿². The results show that the SVM model could predict heterogeneous ignition temperature of coal char particles based on coal proximate analysis accurately. Compared with the Artificial Neural Network (ANN) model, the SVM model is more reasonable and feasible. Besides, a prediction system has been developed by object-oriented high-level language accordingly.