基于G-BP算法的数据挖掘方法研究及应用

Research and application of the data mining method based on G-BP algorithm

将神经网络与遗传算法相结合.提出了G-BP混合算法.该算法在BP神经网络训练过程中,利用遗传算法善于发现最优解区域的特点来优化网络权重值和阈值.在新建项目投资估算的具体应用中,证明该算法克服了传统BP网络算法中的局部极小缺陷,训练速度有很大提高,在数据挖掘中具有实用性.

The hybrid G -BP algorithm was put forward by combining neutral network and genetic algorithm. This algorithm can use genetic algorithm to optimize network weight valve value. The application of the new-constructed project cost estimates showed that this algorithm overcomes the shortcomings of local convergence in the local optimal solution and shortens the training speed. It was of high practicality in data mining.