A Novel Method for Privacy Preservation of Health Data Stream
Ganesh Dagadu Puri · International Journal of Advanced Trends in Computer Science and Engineering · 2020
Most of the conventional privacy preservation methods focus on static datasets.These methods cannot be applied as it is on real-world datasets; which dynamically modify data.If one tuple in the dataset get modified, statistics of complete dataset gets changed.Privacy preservation measures must be applied after modification of dataset.Such re-anonymization of complete dataset is incompetent when large datasets are often changed.When dynamic stream data is considered, we have to apply different privacy techniques which can apply privacy on each tuple.Although several studies have addressed data privacy for static, dynamic and stream data, they are not adequate for avoiding similarity attack and reducing data loss in privacy preservation.Even in the big streaming data privacy preservation, repetition of L-diverse group can take place.This repetition can cause re-identification of individual.Therefore, we identified limitations of data-privacy preservation for stream data and developed more efficient L-diversity algorithm for preserving privacy of data streams.We used hash values of L-diverse groups to find similarity among the groups.Human disease ontology is used to find synonyms of disease terms.Experimental results demonstrated that, our proposed anonymization algorithm reduced data loss of L-diverse group.