大数据中目标信息准确查询仿真研究

摘要

大数据中对目标信息的准确查询,能够为信息分析提供更多的信息资源保障和技术支持.对大数据中目标信息的查询,需要对原始目标信息特征空间进行降维,给出目标信息码本阈值,完成对信息准确的查询.传统方法先给出分布式乘积量化模型的形式化定义,利用训练出的码本模型对分布式数据进行编码,但忽略了对原始数据进行降维处理,导致查询效果不理想.提出基于哈希的大数据中目标信息准确查询方法.提取目标信息的特征,给出信息间的相似性关系,获取两个相关矩阵的归一化的内积,得到基于目标信息字面上的词项的相似度,通过主成分分析理论对原始目标信息特征空间进行降维,利用哈希理论构成了目标信息原始向量的哈希编码,给出训练出的目标信息码本阈值,利用该阈值完成对目标信息准确查询.仿真证明,所提方法查询精度较高,大幅度提升了信息查询的质量.%An accurate query method for target information in big data is propose based on Hash.First of all,the target information feature is extracted and similarly relation in information is given,thus normalization transvection of two related matrices and similarity of lexical item in literal are obtained based on target information,and then dimensionality reduction for feature space of original target information is carried out by theory of principal component analysis and Hash code of original vector of target information is formed via Hash theory to give codebook threshold of target information trained.At last,the accurate query of target information is achieved by utilizing the threshold.The simulation results show that the mentioned method has high query accuracy.It can greatly improve quality of information inquiry.