基于BP人工神经网络的信息过滤技术研究

Research of Information Filtering Technology based on BP ANN

首先介绍了信息过滤中向量空间模型的基本工作原理。针对目前信息过滤技术过滤精度低、速度慢的缺点,利用BP人工神经网络网络稳定、学习能力强的特点,重点研究了BP人工神经网络特征输入数据的取得、网络拓扑结构的选择、BP人工神经网络的训练过程、神经网络产生结果数据的处理方式,在此基础上提出了一种基于BP人工神经网络的信息过滤模型,实验证明,此模型过滤效率有了显著提高。

At first this paper descirbes the basic operation principle of Vector Space Model in information filtering. In view of the shortages of current information filtering technology as low precision and low speed, this paper studied the acquirement of BP ANN input data , the choice of network topology, the training process and the processing method of BP ANN output data by using the advantage of BP ANN such as stable performance and strong learning capability. And in the end, this paper proposes a new information filtering model based on BP ANN. The experiment results proved that this model filter is remarkable in efficiency.