大数据中用户所需信息资源检测仿真

发布时间：2018-04-23 23:13

本文选题：大数据 + 用户所需信息　；参考：《计算机仿真》2017年11期

【摘要】：通过对数据缺失特征进行检测实现信息资源的有效检测,能够保证大数据中用户所需信息的完整性和准确性,对用户所需信息资源的检测,需要计算出数据特征距离,分析分配缺失数据属性特征权重,完成信息资源的检测。传统方法定义约束容差集合差异度,计算出不完备数据特征集合内全部对象的总体相异程度,但忽略了分析分配缺失数据属性特征权重,导致检测精度偏低。提出基于遗传优化的大数据中用户所需信息资源检测方法。结合遗传优化思想估计用户所需信息均值和协方差矩阵,以用户所需信息的对数似然函数作为目标函数,通过已有缺失数据样本获得参数的相应约束条件,建立大数据中缺失数据估计模型,消除用户所需信息量纲的影响,计算出数据特征的距离,利用聚类分析分配缺失数据属性特征权重,完成大数据中用户所需信息资源检测。实验结果表明,所提方法具有较高的填补准确性,且可扩展性较强。
[Abstract]:By detecting the missing features of the data to realize the effective detection of the information resources, the integrity and accuracy of the information needed by the users in big data can be guaranteed, and the distance of the data features should be calculated for the detection of the information resources needed by the users. Analysis and allocation of missing data attribute feature weights to complete the detection of information resources. The traditional method defines the difference degree of constraint tolerance set, calculates the total difference degree of all objects in the incomplete data feature set, but ignores the analysis and allocation of missing data attribute feature weights, which results in low detection accuracy. Based on genetic optimization, a method of information resource detection for users in big data is proposed. Combined with genetic optimization idea to estimate the average value and covariance matrix of information required by the user, the logarithmic likelihood function of the information required by the user is taken as the objective function, and the corresponding constraint conditions of the parameters are obtained through the existing missing data samples. The missing data estimation model in big data is established to eliminate the influence of the information dimension needed by the user, the distance of the data features is calculated, and the attribute weights of missing data are assigned by cluster analysis, and the information resources needed by the user in big data are detected. The experimental results show that the proposed method has higher filling accuracy and better scalability.
【作者单位】：江西应用科技学院信息工程学院;
【分类号】：TP311.13

【相似文献】