基于IPSO一LSTM神经网络的施肥量预测系统

打开文本图片集

中图分类号：S513：TS213.4 文献标识码：A 文章编号：2095-5553（2025）12-0208-07

Abstract：Toachieve precise prediction ofcorn fertilization dosage byusing a neural network andaddress theissueof environmentalpollutioncausedbylowfertlizerutilizationrates，anovelpredictionmodelcombining theImprovedParticle Swarm Optimization （IPSO）algorithm and Long Short-Term Memory（LSTM）neural network，termed IPSO—LSTM， was designed.Climatedata and field management information fromexperimental plots werecollcted.An LSTM neural network was introduced，andan improved particleswarm algorithm was used to optimize the LSTM neural network， constructing the IPSO—LSTMfertilization prediction model.Resultsindicatethat，compared toother network models suchas PSO—LSTM，the IPSO—LSTM-based fertilization prediction system demonstrates beter performance in terms of accuracy and stability. The model achieves a mean absolute percentage error of 1.2% ，a mean squared error of 1.445 （kg/hm2）2 ，a root mean squared error of 1.202kg/hm2 ，and a mean absolute error of 0.968kg/hm2 . The IPSOLSTMmodelefectively improves theprecisionofcornfertilization，contributing totherealizationofprecision management in agricultural production. It has significant practical application value and promotion prospects.

Keywords：corn; fertilizer dosage prediction; particle swarm optimization algorithm; long short-term memory network

0 引言

农业关键技术的发展显得尤为重要[1]。（剩余10761字）