基于卷积神经网络的顶岗实习管理系统数据挖掘研究

Research on data mining of post practice management system based on convolutional neural network

通过数据挖掘方法管理系统数据时,仅依靠关联规则约束,容易导致数据挖掘的泛化误差增大。因此,以顶岗实习管理系统为例,提出基于卷积神经网络的数据挖掘方法。提取顶岗实习管理系统的数据,建立面向主题的数据仓库,结合统计回归分析法和模糊聚类法生成非线性时间序列数据流,采用模糊聚类法设计数据特征提取机制。根据数据特征提取结果分析关联规则,构建卷积神经网络数据挖掘模型,通过特征压缩方法进行数据降维处理,实现挖掘数据的输出。实验结果表明:所提数据挖掘方法与基于决策树和基于一维卷积网络的方法相比,泛化误差较小,能保持在[-0.05,0.05],可以获取更加精确的信息挖掘结果,具有较好的实际应用效果。

When managing the data of system through data mining method,only relying on the constraints of association rules is easy to increase the generalization error of data mining.Therefore,the research on data mining of post practice management system based on convolutional neural network is proposed.Extract the data contained in the post internship management system and establish a subject oriented data warehouse.Combining the statistical regression analysis method and fuzzy clustering method,the nonlinear time series data stream is generated,and the fuzzy clustering algorithm is used to design the data feature extraction mechanism.According to the data feature extraction results,the association rules are analyzed,the convolution neural network data mining model is constructed,and the data dimensionality is reduced through the feature compression method to realize the output of mining data.The experimental results show that the generalization error of the proposed data mining method is lower than that of the method based on decision tree and one-dimensional convolutional network,which can be kept within[-0.05,0.05],and can obtain more accurate information mining results,and has a good practical application effect.