System evolution analytics: data mining and learning of complex and big evolving systems

Abstract

Era of artificial and computational intelligence leads to various kinds of systems that evolves. Usually evolving system contains evolving inter-connected entities (or components) that makes evolving states created over time. Several evolving computing systems are stored and managed in a repository, which makes a kind of time-variant or non-stationary data. Numbers of system repositories are increasing across the globe and temporal databases are increasing in size, which makes system maintenance and evolution a challenging task. In this thesis, our objectives is to model an evolving system state as (Si, ERi, ti), such that the system state Si and the evolution representor ERi is representing system at the ith time point ti, where ‘i’ varies from 1 to N. We expressed an evolving system as a state series, SS = {S1, S2… SN}, such that each state is pre processed to make an evolution representor ER = {ER1, ER2… ERN} for example evolving networks EN = {EN1, EN2… ENN}. This modelling is used to do system evolution analysis. The state series represents a non-linear time-variant evolving system. A state series of an evolving system is stored and managed in a centralized repository. We introduce a System Evolution Analytics (SysEvo-Analytics) based on proposed evolving system mining and evolving system learning, which are techniques for an evolving system represented as a set of temporal data. The evolving (or temporal) networks represented system state series are combined to form an evolution representor that can be used for analysis purpose. Our goal is to analyze the evolving interconnected entities (or features) in a state series. We present following contributory approaches: