Intelligent prediction model based on PAC and RBFNN and its application

Because key technical index in complex industry production process is so difficult to measure directly online, an intelligent prediction model based on PCA and RBFNN is developed. The model consists of four modules: data collection and pretreatment, prediction, online modification and effect estimate. PCA technology reduces the dimensions of secondary variables. Among RBFNN, the nearest neighbor algorithm was used to adjust centers and avoid the disadvantage of the K-means, which relies on the initial central position and is possible to enter local minimum point. The recursive least square method was used to calculate the output weights of middle layers. Industrial application show that the prediction model can reflect the actual operation condition, and solve the problem of the measurement difficulty of the key variables to a great extent, and has high reliability, perfectly precision and good real time character, meet the requirement of real-time control.