Privacy Preservation in eLearning: Exploration and Analysis

Nowadays, a big amount of data is collected in eLearning environments, tracking students’ behavior and their performance of learning activities. Also, a part of educational data is used by third parties for statistical or research purposes. In many cases, the datasets are transferred and processed without any techniques for students’ identity protection and there are possibilities after attacks private and sensitive data to be revealed. The aim of the paper is to present the results from conducted explorations and analysis about applying privacy preserving algorithms k-anonymity and ($\varepsilon,\delta$)-differential privacy on data, collected in eLearning environment. The balance between students’ privacy protection and usefulness of output information is discussed considering several privacy parameters. Machine learning is used to predict the most suitable privacy models and in this way to support the decision making of data holder/owner.