Reframing in Clustering: An Introductory Survey

Abstract

Reframing is an essential task for improving the performance of machine learning and data mining algorithms in the areas where there are context changes between the source and target domains. A major assumption in many reframing algorithms is that the target domain has some labelled data. However, in many real-world applications, this assumption may not hold. For example, we sometimes have a clustering task in one domain of interest, but we only have sufficient source data in another domain of interest, where the latter data may be in a different feature space or follow a different data distribution. Moreover, both source and target data may be unlabelled. In such cases, reframing in clustering, if done successfully, would greatly improve the performance of clustering by avoiding much expensive data labeling efforts. In recent years, reframing in clustering has emerged as a new clustering framework to address this problem. In this paper, we present a review on the state-of-the-art reframing in clustering approaches, and to the best of our knowledge it has never been done in the literature. We give a definition of reframing in clustering. We also explore some potential future issues in this area of research.