Data encryption and information security using privacy preserving data mining techniques

Abstract

Data in form of information is available in multiple formats like numeric, textual, imageries newlineand audio-visual. Knowledge derived from these datasets using several intelligent approaches newlinefalls into the broad category of Data Mining (DM) . The output so derived in form of newlineknowledge patterns from DM processes helps in ascertaining meaningful insights from raw newlinedata. However, deriving information-rich patterns from data is often associated with an newlineinherent risk of invading sensitive niceties. For instance, subtle data implications include newlinegenetic particulars, biometric statistics, ethnical details, religious inclinations, gender newlinespecifics and healthcare records. Supposing sensitive information is exposed to unauthorized newlinethird party hawkers, it leads to data theft and misuse. Henceforth, it is essential to perform newlineDM tasks without forgoing confidential minutiae s. Conversely, it is quite challenging task for newlinetraditional DM algorithms to identify knowledge patterns from non-sensitive data instances. newlineThese techniques consider entire dataset as input for knowledge modeling without bifurcation newlineof sensitive and non-sensitive instances. Henceforward, there is an inherent need to upgrade newlineDM techniques for including privacy specifications prior to knowledge extraction. In this newlinedirection, Privacy Preserving Data Mining (PPDM) approaches play substantial role for newlineconserving delicate information prior to knowledge extraction process. It is essential for newlinePPDM algorithms to safeguard sensitive data without obstructing utility of data or the DM newlineoutput. In this direction, it is imperative to design effective PPDM algorithms for handling newlineprivacy apprehensions in data. newlineThe current research work focuses on designing novel PPDM algorithms based on newlineprinciples of traditional concealment conservation techniques which aids in safeguarding newlinesensitive information followed by knowledge mining processes. In this direction, the present newlineresearch is carried out on three diverse datasets in three phases to discern the