Spss Statistical Analysis Package And Neural Network Applications, The Combination Of Model In Mineral

Drastic market competition and gradually rigorous environment that enterprise producing and managing make the mineral processing plant more and more attach importance to actual application of the modeling fruit based actual production data. Applying these theories and methods to open out complicated connection of actual production data and model has become one of the points that nowadays prospering nations study.Neural network technique establishes the high accuracy variable's non-linear reflecting model with less experiment workload. With neural network BP model, we can establish non-linear model between various indexes and influence factors, which have well effect of forecast. However applied modeling of the mineral processing plant is a complicate and very difficult work, because the relativity of production data, complexity of the non-linear modeling and the absence of technical parameter quantum data of mineral processing plant, which should be used as independent variable of network and so on.In theory we require linear independence among variables used as inputting or outputting of BP net when we model the mineral processing plant with BP net. However it is always definite relativity in inputting variables or outputting variables in modeling of the mineral processing plant, which sets definite obstacle in high quality modeling. On the other hand redundant inputting variables also lead low modeling efficiency. According to the two questions, we can resolve relative problem of inputting and outputting data, reduce scale of neural network at all, advance popular ability of neural network, and highly improve the quality of modeling by comprehensively utilizing Principal Component Analysis of SPSS & BP technique.The application of mineral procession mathematic model based neural network begin to show clue in production of mineral procession, however it is commonly only used to forecast effect of mineral procession, and little guide mineral processing. MEANS of SPSS search possible combination of dosage of huangyao and dosage of 2# oil when the combination of grade of concentrate and recovery is better, which can be used to guide dosage of mineral procession so as to improve benefit of mineral procession.The modeling of mineral processing plant based the combination of Exploratory Data Analysis SPSS and neural network in this article improve quality of modeling and enrich content of neural network; At the same time it make mathematic model of mineral processing plant not only be used to forecast effect but also forecast condition of controlling.