基于深度学习的基层网络数据个性化挖掘算法

Personalized Data Mining Algorithm for Grassroots Network Based on Deep Learning

为提升基层网络数据挖掘精度与效率,有效应用基层网络数据提供帮助,提出基于深度学习的基层网络数据个性化挖掘算法,设计基于模糊神经网络的基层网络数据个性化挖掘算法过程,通过数据准备阶段清洗、选取及转化初始基层网络数据,得到高精度完整统一的待挖掘基层网络数据,划分其为训练组与测试组,构建包含输入层、模糊输入层、隐含层、模糊输出层及期望输出层的五层模糊神经网络,运用训练组基层网络数据训练该模糊神经网络,裁剪掉训练后模糊神经网络内的冗余权值规则,提取出最大权值规则,运用该规则对测试组基层网络数据实施挖掘。实验结果表明,上述算法实际应用中收敛速度较高,在训练与测试速度方面具有较大优势,可实现高精确、高查全及高重合度的精准挖掘,为基层网络数据的有效利用奠定基础。

In order to improve the accuracy and efficiency of grass-roots network data mining and provide help for the effective application of grass-roots network data, this paper proposes a personalized grass-roots network data mining algorithm based on deep learning. A personalized grass-roots network data mining algorithm process was designed based on fuzzy neural network, and the initial grass-roots network data were cleaned, selected and transformed through the data preparation stage. Then we obtained the high-precision, completed and unified basic network data to be mined, divided it into training group and test group, constructed a five-layer fuzzy neural network including input layer, fuzzy input layer, hidden layer, fuzzy output layer and expected output layer, trained the fuzzy neural network with the basic network data of the training group, and cut out the redundant weight rules in the fuzzy neural network after training. The maximum weight rule was extracted and used to mine the grass-roots network data of the test group. The experimental results show that the algorithm has high convergence speed and high precision, high recall and high coincidence.