Dissertation > SCIENCE AND > Journal of Systems Science > Systems Engineering > System technology > Forecasting techniques

Research on Grain Yield Forecasting Based on Gene Expression Programming

Author LiYin
Tutor HeDongJian
School Northwest University of Science and Technology
Course Applied Computer Technology
Keywords Grain Yield Forecasting Forecasting Module Gene Expression Programming Data Mining Genetic Programming
CLC N945.24
Type Master's thesis
Year 2010
Downloads 112
Quotes 0
Download Dissertation

The grain problem is a very important issue with strong relation to the establishment of a harmonious socialist society; furthermore, it also concerns the peaceful development of every country in the world and the full-scale development of the comprehensive progress of human society. The prediction of grain yield not only provides technical support and theoretical foundation of grain production to the local government, but also plays an important role in the promotion of the all-round development of the social economy.With data mining as the guiding ideology, software engineering technology as the core and the forecasting model of gene expression programming as the basis, this article made close study on the grain yield data of Shaanxi Province and developed a modeling emulation platform based on gene expression programming.The main contents in the paper are as follows:(1) Analysis on the factors influencing grain yield prediction. In this thesis, we employed fuzzy optimum seeking theory to adopt six factors, including the preference matrix and weighted value of the indexes, the owning quantity of farm machines, rural electric power, fertilizer, acreage of cultivated farmland, per capita area of cultivated farmland in every village, and grain acreage, as references to predict the grain yield. It is proved from validation that the method can provide accurate grain yield prediction.(2) We compared various models of grain yield prediction dependently, and we employed the model of gene expression programming as the model of grain yield prediction for the first time in China. According to the specific features of grain yield prediction, three models of grain yield prediction, BP neural network, support vector machine and gene expression programming, have been applied in the research of grain yield prediction. The result of experiment shows that gene expression programming has a logical individual evolution, a quick evolving speed, a high success rate, and has excellent performance for grain yield prediction.(3) We developed a modeling emulation platform based on gene expression programming approach. Based on the mathematical model of gene expression programming, this platform is implemented by VC and MATLAB programming. And in our system, the user requirement of various disciplines and domains is integrated to achieve the corresponding forecasting tasks.(4) It is demonstrated from the analysis and corroboration on the datum of different subject areas that the simulation platform can quickly realize the comparison experiment between the measured value and the predicted value, thus obtain the simulation results of relevant datum. In all, this technology has the practical utility of scientific prediction with high accuracy.

Related Dissertations
More Dissertations