Communication theoretic prediction on networked data

Prediction based on observed data is one of the major purposes in (big) data analytics, and has shown great impacts in many applications, including engineering, social science, and medical treatments. Statistical machine learning has been widely adopted to deal with such problem. In this paper, we analogize the relationship among data variables as a sort of generalized social network [1], that is, networked data. Consequently, a direct causal relationship from one data variable to another is thus equivalent to information transfer over a communication channel. Prediction based on data variables is consequently to maximize utilizations of information conveyed over communication channels. Therefore, we introduce the concept of adaptive equalization to data analytics in this paper, which allows us to select appropriate data variables and optimum depth of observations for prediction. We illustrate by finance market data to show surprisingly good performance using this simple methodology. This result not only indicates a new direction to knowledge discovery and inference in big networked data analytics based on communication theory, but also shows the consistency with the newly developed information coupling.