Grant: Engage Grants Program

ISM Canada requires research on privacy protocols and techniques for privacy preserving data mining becauseof anticipated opportunities in mining government data. A particular concern is about what happens toconfidential data when it is sent to a cloud-based service for data mining. ISM would like to know whatguarantees about the privacy of such data can be made if a new technique is developed that encrypts data beforeit is sent to the data mining service. Since the provider of the data mining service has access to the data, amethod of keeping the data and the mined patterns confidential is required.The main objective of the research project is to design and evaluate an encryption scheme (protocol) with aguaranteed formal privacy in which (i) the data owner encrypts its data using an encryption module, (ii) theservice provider conducts data mining on the encrypted data and sends the results to the owner, and (iii) adecryption module translates the results and recovers the true identity of the returned patterns and associationrules. We aim to apply the idea of a blind digital signature in our scheme, which is used in protocols where thesigner and message author are different parties (e.g., election systems and digital cash schemes).In the evaluation part of the research, first the scheme will be implemented (e.g., using C++ and SQL Server)and verified using some artificial data. Next, we aim to produce a software prototype and conduct anexperimental analysis of our scheme using a large real data set by comparing our results with the results fromcommercial software packages, such as IBM Cognos and IBM Identity Insight.