Understanding Technological Abuse: An Exploration of Creepware

Abstract

Prior research has explored Intimate Partner Surveillance (IPS) apps, and developed a machine learning model to detect these apps on victim’s phones. Our exploration bolsters these detection systems by providing insights on a dataset of millions of devices and their installed apps. Moreover, we provide descriptions and patterns of the “creepware” ecosystem that is currently present in the app marketplace. With a taxonomy of the creepware space, we attempt to build a machine learning model to label the over three million apps that exist in our dataset. This would have helped in the development of a feature space to classify devices based on their theoretical use cases. However, due to a lack of data and issues with data governance, we instead present an analysis of the previously made taxonomy using LDA, as well as an attempt to cluster devices using a feature space created out of the current existing taxonomy.