The Application of Data Mining Algorithm in Grain Output Prediction

Abstract

Genetic algorithm and neural network are two basic methods of data mining. This paper studies the related theory of genetic algorithm and neural network and, on this basis, proposes an improved mining method basing on the combination of genetic algorithm and LM optimization algorithm. This method uses genetic algorithm in two stages to improve the quality of network training. It first through the genetic algorithm to get an approximation of the overall solution by coarse control, as the initial value, and then takes genetic algorithm and LM optimized neural network algorithm to alternately train the network. By an example calculation, we can see it is feasible for crop output prediction model basing on neural network to finally inquire of the future crop output by fitting the historical data. The model achieves good results through the optimization of genetic algorithm and LM.