云计算环境下的大数据特征挖掘技术研究.docx

云计算环境下的大数据特征挖掘技术研究 摘 要： 基于梯度采样的大数据特征挖掘方法在大数据挖掘过程中，逻辑性较差，仅从单层角度进行大数据特征挖掘，极大降低大数据特征挖掘的精度。对此，提出基于云计算环境下大数据特征挖掘模型的特征挖掘方法，其大数据存储体系层包括多源信息资源服务层、核心技术层、多源信息资源平台服务层以及多源信息资源基础层，采用大数据特征分布函数实施大数据特征匹配，实现大数据特征初步挖掘；大数据特征挖掘处理层对存储体系层中的大数据进行抽取、转换、清洗、集成与加载实现大数据预处理，采用FP?tree对预处理后的大数据进行准确特征挖掘，通过用户层将获取的大数据特征挖掘结果反馈给用户，完成云计算环境下大数据特征挖掘。实验结果表明，所提方法在云计算环境下大数据特征挖掘方面准确率较高，具有低能量开销的优势。 关键词： 云计算； 梯度采样； 大数据； 存储体系； 特征挖掘； 能量开销 中图分类号： TN919?34； TP391.4 文献标识码： A 文章编号： 1004?373X（2018）20?0161?04 Abstract： The big data feature mining method based on gradient sampling has poor logicality during the process of big data mining， and conducts big data feature mining only from the single layer perspective， which greatly reduces the precision of big data feature mining. Therefore， a feature mining method based on the cloud computing environment is proposed for the big data feature mining model. The big data storage system layer consists of the multi?source information resource service layer， core technology layer， multi?source information resource platform service layer， and multi?source information resource base layer. The big data feature distribution function is used to implement big data feature matching and realize initial mining of big data features. In the big data feature mining processing layer， extraction， transformation， cleaning， integration and loading of big data in the storage system layer are conducted to realize big data preprocessing. The FP?tree is used to conduct accurate feature mining of preprocessed data. The obtained big data feature mining results are fed back to users by means of the user layer， so as to complete big data feature mining in the cloud computing environment. The experimental results show that the proposed method has high accuracy of big data feature mining in the cloud computing environment， and has the advantage of low energy overhead. Keywords： cloud computing； gradient sampling； big data； storage system； feature mining； energy overhead 0 引 言 云计算平台具有高效分配动态资源、根据用户请求生成动态计算