一种基于BP神经网络的离散化方法

摘要

With the further development of artificial intelligence and data mining technology, big data gradually into view, in the process large data, discrete processing is an essential link . In this paper, in the learning process by introducing the momentum BP neural network learning method to improve the stability and accuracy of BP neural network , reducing the learning error BP neural network , and on this basis, proposes a BP neural network discretization method to achieve a discrete handling of continuous attri⁃butes . Algorithm analysis and experiments show that the algorithm is feasible.%随着人工智能和数据挖掘技术的深入发展，大数据逐步进入人们的视野，在大数据的处理过程中，离散化处理是一个必不可少的环节。本文通过在BP神经网络的学习过程中引入动量学习法，进一步完善了BP神经网络在学习方面的局限性，降低了BP神经网络的训练误差，在此基础上提出了一种基于BP神经网络的离散化方法，实现了对连续属性的离散化处理。算法分析和实验证明，本算法是切实可行的。