Educational data mining has allowed for large improvements in educational outcomes and understanding of educational processes. However, there remains a constant tension between educational data mining advances and protecting student privacy while using educational datasets. Publicly available datasets have facilitated numerous research projects while striving to preserve student privacy via strict anonymization protocols (e.g., k-anonymity); however, little is known about the relationship between anonymization and utility of educational datasets for downstream educational data mining tasks, nor how anonymization processes might be improved for such tasks. We provide a framework for strictly anonymizing educational datasets with a focus on improving downstream performance in common tasks such as student outcome prediction. We evaluate our anonymization framework on five diverse educational datasets with machine learning-based downstream task examples to demonstrate both the effect of anonymization and our means to improve it. Our method improves downstream machine learning accuracy versus baseline data anonymization by 30.59%, on average, by guiding the anonymization process toward strategies that anonymize the least important information while leaving the most valuable information intact.
more »
« less
PMDG: Privacy for Multi-perspective Process Mining Through Data Generalization
Anonymization of event logs facilitates process mining while protecting sensitive information of process stakeholders. Existing techniques, however, focus on the privatization of the control-flow. Other process perspectives, such as roles, resources, and objects are neglected or subject to randomization, which breaks the dependencies between the perspectives. Hence, existing techniques are not suited for advanced process mining tasks, e.g., social network mining or predictive monitoring . To address this gap, we propose PMDG, a framework to ensure privacy for multi-perspective process mining through data generalization. It provides group-based privacy guarantees for an event log, while preserving the characteristic dependencies between the control-flow and further process perspectives. Unlike existing privatization techniques that rely on data suppression or noise insertion, PMDG adopts data generalization: a technique where the activities and attribute values referenced in events are generalized into more abstract ones, to obtain equivalence classes that are sufficiently large from a privacy point of view. We demonstrate empirically that PMDG outperforms state-of-the-art anonymization techniques, when mining handovers and predicting outcomes.
more »
« less
- Award ID(s):
- 1952225
- PAR ID:
- 10471888
- Publisher / Repository:
- Springer-Verlag
- Date Published:
- Journal Name:
- 35th International Conference on Advanced Information Systems Engineering (CAiSE) 2023
- Format(s):
- Medium: X
- Location:
- Zaragoza, Sapin
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Process mining is a technique for extracting process models from event logs. Event logs contain abundant information related to an event such as the timestamp of the event, the actions that triggers the event, etc. Much of existing process mining research has been focused on discoveries of process models behind event logs. How to uncover the timing constraints from event logs that are associated with the discovered process models is not well-studied. In this paper, we present an approach that extends existing process mining techniques to not only mine but also integrate timing constraints with process models discovered and constructed by existing process mining algorithms. The approach contains three major steps, i.e., first, for a given process model constructed by an existing process mining algorithm and represented as a workflow net, extract a time dependent set for each transition in the workflow net model. Second, based on the time dependent sets, develop an algorithm to extract timing constraints from event logs for each transition in the model. Third, extend the original workflow net into a time Petri net where the discovered timing constraints are associated with their corresponding transitions. A real-life road traffic fine management process scenario is used as a case study to show how timing constraints in the fine management process can be discovered from event logs with our approach.more » « less
-
Research and practical development of data-anonymization techniques have proliferated in recent years. Yet, limited attention has been paid to examine the potentially disparate impact of privacy protection on underprivileged subpopulations. This study is one of the first attempts to examine the extent to which data anonymization could mask the gross statistical disparities between subpopulations in the data. We first describe two common mechanisms of data anonymization and two prevalent types of statistical evidence for disparity. Then, we develop conceptual foundation and mathematical formalism demonstrating that the two data-anonymization mechanisms have distinctive impacts on the identifiability of disparity, which also varies based on its statistical operationalization. After validating our findings with empirical evidence, we discuss the business and policy implications, highlighting the need for firms and policy makers to balance between the protection of privacy and the recognition/rectification of disparate impact. This paper was accepted by Chris Forman, information systems.more » « less
-
In critical infrastructure (CI) sectors such as emergency management or healthcare, researchers can analyze and detect useful patterns in data and help emergency management personnel efficaciously allocate limited resources or detect epidemiology spread patterns. However, all of this data contains personally identifiable information (PII) that needs to be safeguarded for legal and ethical reasons. Traditional techniques for safeguarding, such as anonymization, have shown to be ineffective. Differential privacy is a technique that supports individual privacy while allowing the analysis of datasets for societal benefit. This paper motivates the use of differential privacy to answer a wide range of queries about CI data containing PII with better privacy guarantees than is possible with traditional techniques. Moreover, it introduces a new technique based on Multipleattribute Workload Partitioning, which does not depend on the nature of the underlying dataset and provides better protection for privacy than current differential privacy approaches.more » « less
-
Obtaining an accurate specification of the access control policy enforced by an application is essential in ensuring that it meets our security/privacy expectations. This is especially important as many of real-world applications handle a large amount and variety of data objects that may have different applicable policies. We investigate the problem of automated learning of access control policies from web applications. The existing research on mining access control policies has mainly focused on developing algorithms for inferring correct and concise policies from low-level authorization information. However, little has been done in terms of systematically gathering the low-level authorization data and applications' data models that are prerequisite to such a mining process. In this paper, we propose a novel black-box approach to inferring those prerequisites and discuss our initial observations on employing such a framework in learning policies from real-world web applications.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Conference Proceeding:
- https://doi.org/10.1007/978-3-031-34560-9_30
