The Temperature Prediction in Blast Furnace Base on Fuzzy Fuzzy Clustering Analysis and Least Squares Support Vector Machine

For the blast furnace temperature forecast in iron and steel enterprises, this paper divides the train samples into several categories basing on fuzzy C-means clustering (FCM). Then we train the samples by least squares support vector machine (LSSVM). According the result of cluster analysis, we classify the predict samples. This improved method can avoid the shortcomings of training the samples blindly and improve prediction accuracy of the LSSVM model effectively. According the result of simulation, this method has a significant improvement in the trends and the accuracy of temperature forecasting. It also can play a guiding role in the blast furnace management operations.