【摘要】 对大数据网络中需求目标干扰区域进行有效过滤,可去除大数据网络中的冗余数据,提高网络需求目标检索的效率。进行需求目标干扰区域过滤时,不同需求目标干扰区域具有不同过滤特征,导致利用传统方法进行需求目标干扰区域过滤时,需要根据不同的目标设定不同的过滤阈值,降低了需求目标干扰区域过滤效率。提出一种混合径向基函数插值和贝叶斯分类的大数据网络中需求目标干扰区域过滤方法。上述算法首先采用大数据网络中不同的数据样本点与残缺样本点距离以及标签分类进行径向基函数插值,利用贝叶斯分类进行需求目标干扰区域数据过滤,通过比较大数据网络中数据样本的干扰数据和非干扰数据的先验概率、特征项的类条件概率预测未知数据某一类型的后验概率,最终依据后验概率大小判断是否为干扰区域,对其进行去除,完成大数据网络中需求目标干扰区域过滤。仿真结果表明,所提算法进行大数据网络中需求目标需求目标干扰区域过滤精准性高,过滤效率较好。更多还原

【Abstract】 Demand for large data network target effectively filtering interference area,can eliminate the redundant data in the large data network,improve the efficiency of network demand target retrieval. Demand goal interference filter of the region,the different interference filter area with different characteristics,and lead to the use of traditional method for demand goal interference filter of the region,according to different target set different threshold filtering,reduced the demand for filtration efficiency target jamming area. Put forward a hybrid radial basis function interpolation and bayesian classification of big data in the network demand target jamming area filtering methods. The algorithm firstly using data from a large data network in different distance between sample points and the imperfect sample points and the tag on the radial basis function interpolation,using bayesian classification to interfere with the regional data filter,by comparing the data samples in the big data network data and the interference of prior probability,characteristics of conditional probability forecast the unknown data a certain type of posterior probability,finally on the basis of a posteriori probability area size to determine whether to interference,to remove the complete big data in the network demand goal interference filtering area. Simulation results show that the proposed algorithm is big data in the network demand target jamming area high filtration precision,good filtration efficiency.更多还原