Coal Security Forewarning Based on PCA-AR and K-means Clustering

Based on the factors including the supply and demand,transportation,cataclysm,economy and environment in coal production,the forewarning index system of coal security with 5 major categories and 12 indexes was established.By utilizing principal component analysis,autoregression and K-means clustering,a model for coal security forewarning was proposed,which can predict the degree of coal security and its categories,i.e.risk,attention or safety.The results by applying the model in coal security forewarning showed that the coal energy would be safe in China in 2007.