科技通报
科技通報
과기통보
Bulletin of Science and Technology
2015年
10期
154-156
,共3页
贝叶斯粗糙集%云数据%融合
貝葉斯粗糙集%雲數據%融閤
패협사조조집%운수거%융합
Bayesian Rough Set%cloud data%fusion
云计算环境下,需要对云数据特征进行深度融合,提高对云数据的调度和决策能力.传统的云数据融合算法采用置信增益概率分配算法,当云数据出现多重特征时,融合深度不够,信息提取效果不好.提出一种基于贝叶斯粗糙集的云数据深度融合算法.引入了置信增益函数贝叶斯粗糙集,得到贝叶斯粗糙集云数据模型构建,在特征空间关系中进行特征合并,进行决策表决策属性分区处理,提高融合精度,依据信任函数最大化原则确定新对象的决策属性取值,实现云数据深度融合算法改进.仿真实验表明,采用该算法,能有效提高数据融合深度和精度,稳健性较好,可以明显的抑制噪声的影响,并提高20 dB左右的特征空间增益,算法在高维空间中仍体现出了较为明显的数据融合优势,该算法在云计算和云数据信息处理等领域具有较好应用前景.
雲計算環境下,需要對雲數據特徵進行深度融閤,提高對雲數據的調度和決策能力.傳統的雲數據融閤算法採用置信增益概率分配算法,噹雲數據齣現多重特徵時,融閤深度不夠,信息提取效果不好.提齣一種基于貝葉斯粗糙集的雲數據深度融閤算法.引入瞭置信增益函數貝葉斯粗糙集,得到貝葉斯粗糙集雲數據模型構建,在特徵空間關繫中進行特徵閤併,進行決策錶決策屬性分區處理,提高融閤精度,依據信任函數最大化原則確定新對象的決策屬性取值,實現雲數據深度融閤算法改進.倣真實驗錶明,採用該算法,能有效提高數據融閤深度和精度,穩健性較好,可以明顯的抑製譟聲的影響,併提高20 dB左右的特徵空間增益,算法在高維空間中仍體現齣瞭較為明顯的數據融閤優勢,該算法在雲計算和雲數據信息處理等領域具有較好應用前景.
운계산배경하,수요대운수거특정진행심도융합,제고대운수거적조도화결책능력.전통적운수거융합산법채용치신증익개솔분배산법,당운수거출현다중특정시,융합심도불구,신식제취효과불호.제출일충기우패협사조조집적운수거심도융합산법.인입료치신증익함수패협사조조집,득도패협사조조집운수거모형구건,재특정공간관계중진행특정합병,진행결책표결책속성분구처리,제고융합정도,의거신임함수최대화원칙학정신대상적결책속성취치,실현운수거심도융합산법개진.방진실험표명,채용해산법,능유효제고수거융합심도화정도,은건성교호,가이명현적억제조성적영향,병제고20 dB좌우적특정공간증익,산법재고유공간중잉체현출료교위명현적수거융합우세,해산법재운계산화운수거신식처리등영역구유교호응용전경.
Computing environment, the need for cloud data characteristics of the depth of integration, improve the cloud da-ta scheduling and decision making ability. Traditional data fusion use confidence gain probability assignment algorithm cloud, when multiple feature fusion cloud data, not enough depth, the effect of information extraction is not good. This paper proposed a fusion method for cloud data depth based on Bayesian Rough sets. Introduce a confidence gain function Bayes-ian Rough set, obtained the Bayesian Rough Set cloud data model, feature combination in the spatial relationship feature, decision table decision attribute partition processing, to improve the fusion accuracy synthesis results after data fusion based on trust function maximum principle to determine the decision attribute values the new object, implementation of cloud data depth improvement fusion algorithm. Simulation results show that using this algorithm can effectively improve the data fusion depth and precision, good stability, can suppress the noise effect obviously, and improve the feature space gain about 20 dB algorithm in high dimension space, still reflects the obvious advantages of data fusion, the algorithm will be in the cloud computing and cloud data information processing and other fields wide application prospect.