Distributed Data Platform for Machine Learning Using the Fog Computing Model

In this paper, we propose a machine learning-driven information platform for distributed data management without data accumulation using a fog computing model. It helps in analyzing the features of managed contents through several kinds of methods and their synchronization among the distributed nodes. Moreover, the feature model generated based on the features of all contents is achieved by combining the respective feature models distributed to some nodes. In other words, a flexible feature model is created by the combining feature models adapting to targets arbitrarily. In this paper, we propose the basic ideas and some basic functions for the data management platform among distributed nodes. Our evaluation confirms the effectiveness of combining the feature models using datasets.