Research On BP Neural Network Based On Factor Analysis And Its Application In Rational Synthesis Of Microporous Materials

Data mining technology can supply powerful support and reliable assurance for right business decision and solve the contradiction between rapidly incremental data and lagging data analysis methods. Data mining is the process of discovering interesting knowledge, such as patterns, associations, changes, anomalies and significant structures, from large amounts of data stored in databases, data warehouses, or other information repositories. It includes four parts: data preparation, mining, pattern evaluation and knowledge representation. Data mining chiefly use class description, association, classification, prediction, clustering and time-series analysis. Also, data mining is a young interdisciplinary field, drawing from areas such as traditional machining learning, neural network and statistics. At the same time, the development of database technology and the application of database technology are very useful to high performance data mining.As an old experiment science, chemistry involves a mass of reaction information, which is hard to be analyzed by traditional measure however can be handled by computer's huge memory and high speed computational ability. Much study has been focused on the exploration of new microporous materials for their widely applications in the field of catalysis, environment and functional material. However, microporous materials are still prepared by traditional try-and-trial method, which is always fruitless and aimless. In order to meet the great need for the materials with interesting properties, molecular engineering directed by function occurs and becomes more and more important with the development of science technique. In spite of great effortsthe formation mechanism of microporous and related materials are still unclear. Therefore, it is very difficult to achieve a rational synthesis of target materials which are so significant and valuable.Data mining can discover valuable information from a large amount of data. As a data mining technique,BP neural network is an important tool in the area of chemistry.Considering its shortage, this work presents a new mode of data mining process. The pretreatment with factor analysis was used to reduce the redundant attributes. By factor analysis we can find few random variables which can control all the variables and then describe the correlation of multiple variables through analysis variables'inner structure of correlative coefficient matrix. Based on these steps,the results show that application of the BP neural network that was built has good predicting capacity. This work will further assist in rational synthesis of microporous materials. This full text divides into five chapters altogether:Chapter one is the introduction. This chapter states the applications of the data mining technique and the development of the molecular engineering. Through introducing the situation of this field in the world at present, we point out the source of the thesis. Finally, we will introduce the content and structure of this text.Chapter two is the research on BP neural network and factor analysis. Through comparing the past web technology, we come to a conclusion that the pretreatment with factor analysis have greater superiority and better prospect in the aspect of improving the performance of the BP neural network used to analysis the synthesis data of the aluminophosphate zeolite.The following chapters introduce the whole process of the FABP Systemfrom background, requirement analysis, and design, coding and testing. Chapter three mainly analyses the data of the aluminophosphate zeolite and pretreat it with the technique of factor analysis. Then we build the prediction model of the special chemical product using BP neural network.Chapter four builds other three prediction models to compare the performance of the FABP. This chapter is the introduction of testing process used in the analysis of the real data. We analyze the superiority of the FABP system from the performance and accuracy.Chapter five summarizes full text. This chapter points out the innovation of this text, and put forward the main research direction of our work in the future.