中文摘要：

为提高产品销售预测的准确性,为企业生产决策提供有力的参考依据,建立BP_Adaboost和计划评审技术PERT混合销售预测模型。将BP神经网络模型与Adaboost算法进行结合,克服单纯使用BP神经网络容易陷入局部极小值的问题;计划评审技术PERT有效利用销售管理人员的经验,在一定程度上实现对销售数据偶然性的预测;将BP_Adaboost和计划评审技术PERT组成混合模型进行销售预测。实验结果表明,该模型提高了销售预测的准确性和有效性。

英文摘要：

To improve the accuracy of sales prediction and provide a powerful reference for decision-making for the enterprise production,a hybrid sales prediction model which combined BP_Adaboost with PERT was established.First,the BP（back propagation）neural network model was combined with Adaboost algorithm,which overcame the problem that the pure use of BP neural network is easy to fall into local minimum value.Then,the program evaluation and review technique（PERT）used the sales management experience effectively to achieve the prediction of sales contingency to certain extent.Finally,as a hybrid prediction model,the combination of BP_Adaboost with PERT was used to predict sales.The experimental results show that the hybrid sales predicting method improves the accuracy and effectiveness of sales predicting.