基于对称矩阵的云计算和信息共享预测分类器

Symmetric matrix-based predictive classifier for computation and information sharing in cloud

为有效快速地提取和识别数据,使其传输过程简单化,提出基于平行对称矩阵的贝叶斯预测分类器(PSM-PBC),开发高效的大数据计算和信息共享的云环境。并行构建分布式大数据三角对称矩阵,利用Householder变换提升数据运算速率;为提高预测率,使用交叉验证贝叶斯分类器对实值对角数据搜索进行评估。利用贝叶斯类的MapReduce函数对大数据进行有效的预测分析,通过国家电网实际大数据集验证了该方法的准确性和有效性。

To extract and indentify the existing big data effectively,and make the transmission process simplified,Bayes predictive classifier based on parallel symmetric matrix（PSM-PBC）was proposed,which performed efficient data calculation and information sharing in the cloud environment.The distributed large data tridiagonal symmetric matrix was constructed in parallel,and the data operation rate was improved using Householder transformation.To improve the predictive rate,the cross-validated Bayes classifier was used to evaluate the real-valued diagonal data search.The Bayes MapReduce function was used to predict the large data,and the accuracy and validity of the design were verified by the actual large data of the state grid.