【摘要】 热带果树种植适宜度分析对于发展热带果树生产、减少灾害影响具有重要意义。针对统计方法自适应能力低的问题,提出一种将改进的遗传算法(GA)和误差反向传播(BP)神经网络相结合对热带果树种植适宜度进行评价的方法(GA-BP神经网络方法)。首先采用常用的自适应算法对GA的交叉概率和变异概率进行改进,再通过GA得到优化的BP神经网络的初始权值和阈值,在此基础上BP神经网络进行进一步的学习,得到满足误差要求的解。将GA-BP神经网络与传统BP神经网络在twonorm数据集上进行了比较测试,并使用实际气象数据进行了热带果树种植适宜度分析。实验结果表明,GA-BP神经网络的分类准确率较传统BP神经网络高4%左右,网络训练时间减少了3个轮次左右。该方法对热带果树种植适宜度的分析和评价具有应用推广价值。更多还原

【Abstract】 The planting suitability analysis of tropical fruit trees is of great significance for the development of tropical fruit trees and the reducion of disaster effect. Aiming at the problem of low adaptability of statistical methods, a method for the suitability evaluation of tropical fruit trees planting was proposed by combining an improved Genetic Algorithm(GA) and error Back Propagation(BP) neural network(called GA-BP neural network method). Firstly the frequently used adaptive algorithm was used to improve the crossover probability and mutation probability of GA. Then the initial weights and thresholds of BP neural network were optimized by GA. On this basis, BP neural network further studied to obtain the solution satisfying the error requirement. GA-BP neural network was compared with the traditional BP neural network on dataset twonorm, and the planting suitability analysis of tropical fruit trees was carried out based on actual meteorological data. The experimental results show that the classification accuracy of GA-BP neural network is about 4 percentage points higher than that of traditional BP neural network, and network training time decreases about 3 steps. This method has application and promotional value for the analysis and evaluation of planting suitability of tropical fruit trees.更多还原