Performance Analysis of Various Anonymization Techniques for Privacy Preservation of Sensitive Data

Various domains like health care, observing the life of the wild in the forest area, surveillance, etc. have started using the technologies like Wireless Sensor Networks (WSN). Internet of Things (IoT) is also another technology which is used in this regard. WSN based applications generate humungous data due to the continuous monitoring and data logging feature of WSN. This invokes the need to handle humungous magnitude of data and various Big Data frameworks help to handle it. Along with the management of huge amount of data, another issue to be handled is the management of sensitive data. The issue of protecting the secrecy of the data becomes an essentially important area for research. Most of the times, the challenge involved in handling sensitive data is the necessity that the data needs to be published to public media. There exists various problems in data publishing. The problem associated is to preserve the information that is confidential and delicate yet publish useful data. The concept of protecting the privacy of the data involves publishing data without allowing the unauthorized users to access sensitive information. The proposed work focuses on analysis of various techniques in privacy preservation of sensitive data and it is referred as Anonymization. The different anonymization techniques considered for analysis are t-closeness, differential privacy, slicing, k-anonymity and l-diversity. An experimental setup is developed and datasets from Google Financial is used to fractionate the execution of the various anonymization techniques.