Abstract: In order to prevent cloud services from illegally acquiring user’s personal sensitive privacy information,this paper proposed a privacy information publishing detection and protection method for big data.Firstly,the user’s privacy data are classified,and the similarity and the disclosure cost of privacy data are measured respectively.Secondly,accor-ding to the similarity and the disclosure cost,this method detectes whether privacy data required by cloud services contain disclosure chain and key privacy data.Thirdly,continuous privacy datasets,including disclosure chains and key privacy data are decomposed.At the same time,discrete datasets are prevented from being composed into continuous datasets,which do not contain disclosure chains and key privacy data.At last,privacy data chains between discrete privacy datasets and original privacy datasets are discovered by experiment.In terms of precision and recall,the precision of Exact filter is 57% lower than that of non-discrete data sets,while the recall rate is less than 17%.Therefore,the proposed method achieves the purpose of protecting user’s sensitivity privacy information.

Key words: Big data, Key privacy data, Privacy disclosure chain, Privacy enhancement, Privacy release detection