An attribute-based data privacy classification through the Bayesian theorem to raise awareness in public data sharing activity

The growth of the digital era with diverse existing electronic platforms offers information sharing and leads to the realization of a culture of knowledge. Vast amounts of data and information can be reached anywhere at any time, fingertips away. These data are public because people are willing to share them on digital platforms like social media. It should be noted that not all information is supposed to be made public; some is supposed to be kept private or confidential. However, people always misunderstand and are misled about which data needs to be secured and which can be shared. We proposed an attribute-based data privacy classification model using a Naïve Bayesian classifier in this work. It aims to identify and classify metadata (attributes) commonly accessible on digital platforms. We classified the attributes that had been collected into three privacy classes. Each class represents a level of data privacy in terms of its risk of breach. The public (respondent) is determined according to different ages to gather their perspective on the unclassified attribute data. The input from the survey is then used in the Naïve Bayesian classifier to formulate data weights. Then, the sorted privacy data in the class is sent back to the respondent to get their agreement on the class of attributes. We compare our approach with another classifier approach. The result shows fewer conflicting reactions from the respondents to our approach. This study could make the public aware of the importance of disclosing their information on open digital platforms.

Actions (login required)

View Item