中文摘要：

针对稀土期货交易价格的时间序列非平稳的特点,以及向前7步预测的稀土期货交易问题,构建了结合遗传算法和神经网络的混合模型。模型利用遗传算法对神经网络的结构和参数进行优化,克服神经网络弱全局搜索能力和易陷入局部最优值的缺陷,提高预测的逼近精度,为投资者的择时投资提供更准确的参考。

英文摘要：

According to the characteristics of the time series of rare earth futures price and the seven-step-forward prediction of rare earth futures trading,a hybrid model which combines genetic algorithm and the neural network is constructed. The structure and parameters of the neural network are optimized by using genetic algorithm in the model,overcoming the weakness in global search and local optimum,and improving the approximation accuracy of prediction. This research could provide more accurate guidance for time selection of investors.