基于遗传规划法的农村劳动力迁移预测

Application of Genetic Programming to Prediction of Rural Labor Migration

准确预测劳动力迁移量是正确制定迁移决策、合理利用和有序转移农村劳动力的重要依据。由于劳动力迁移量受社会经济中诸多不确定因素影响,而遗传规划终结点集中的随机常数ρ能有效平衡被测量相关因素的影响,因此尝试将遗传规划应用到劳动力迁移预测中。通过训练样本对程序进行训练,建立了受时间变量影响的劳动力迁移预测模型,并由检验样本对模型进行了检验。结果表明:采用遗传规划搜索到的函数具有良好的拟合和预测效果,而且模型结构简单,可有效避免因多种不确定因素影响而造成的人为误差,将其用于劳动力迁移的预测中是可行的。

Accurate prediction of labor migration account is important basis of correct migration decision - making, and reasonable utilization and orderly transformation of rural labor. Because labor migration account is influenced by various uncertain factors of society and economy, but random constant pin terminal sets of GP can balance the influence of related factor of measurand, therefore GP is applied to predict rural labor migration. After the program trained by training sam-ples, prediction model of rural labor migration is affected by time variable is established and is tested by test samples. The results showed that the function which adopt GP has good fitting and forecasting effect, and model with simple structure can effectively avoid artificial error which is formed by various uncertain factors, so it is feasible to predict of rural labor migration.