The quality of learned features by representation learning determines the\nperformance of learning algorithms and the related application tasks (such as\nhigh-dimensional data clustering). As a relatively new paradigm for\nrepresentation learning, Concept Factorization (CF) has attracted a great deal\nof interests in the areas of machine learning and data mining for over a\ndecade. Lots of effective CF based methods have been proposed based on\ndifferent perspectives and properties, but note that it still remains not easy\nto grasp the essential connections and figure out the underlying explanatory\nfactors from exiting studies. In this paper, we therefore survey the recent\nadvances on CF methodologies and the potential benchmarks by categorizing and\nsummarizing the current methods. Specifically, we first re-view the root CF\nmethod, and then explore the advancement of CF-based representation learning\nranging from shallow to deep/multilayer cases. We also introduce the potential\napplication areas of CF-based methods. Finally, we point out some future\ndirections for studying the CF-based representation learning. Overall, this\nsurvey provides an insightful overview of both theoretical basis and current\ndevelopments in the field of CF, which can also help the interested researchers\nto understand the current trends of CF and find the most appropriate CF\ntechniques to deal with particular applications.\n