A Many-Objective Optimization Based Intelligent High Performance Data Processing Model for Cyber-Physical-Social Systems

Abstract

As an emerging paradigm, Cyber-Physical Social System (CPSS) can provide humans with efficient, convenient and personalized services. However, due to the explosive growth of user data, massive and high-dimensional data has become a key issue for CPSS services. In the past, researchers mainly focused on big data analysis methods, ignoring the processing of high-dimensional data before analysis. In response to above problem, this paper focus on the redundant attributes in high-dimensional data. By using intelligent optimization technology, the vision model is optimized for four conflicting goals at the same time, so as to better solve the problem. In order to support this model, a collaboration mechanism (MaOEA-CM) was designed. The algorithm uses the combination of sum of objectives and shift-based density estimation ( <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">${I}_{SDE} {+}$</tex-math></inline-formula> ) indicator as the criterion of gradual co-evolution, and guides individuals to search in the high convergence and multiple directions of CPSS. Experiments on four data sets show that the proposed algorithm and model can significantly improve the accuracy of data processing in CPSS and have higher performance than other feature selection models.