Aerospace production systems, publically traded securities, and countless other systems generate data in time-series formats. The capability to predict future values and outcomes allow optimal decisions and process adjustments to mitigate risk and achieve objectives. This is an application paper that explores improving the accuracy and precision of generating predicted values and decisions with time-series data by integrating existing data mining technologies and information systems. Existing systems are integrated into a System-of-System (SoS) meta-architecture utilizing the Flexible and Intelligent Learning Architecture for SoS (FILA-SoS) [2]. The Overall Objective of the SoS is to maximize the Key Performance Attributes (KPA): Performance of the Predicted Value, Performance of the Predicted Decision, Affordability, Scalabihty, and Robustness. Architectures are generated, assessed, and selected using evolutionary algorithms integrated with a Fuzzy Inference System. The SoS is evaluated with time-series data of publicly traded securities [1]. The results obtained suggest the best or near optimal SoS meta-architecture to improve predictions and decisions of time-series data versus single or hybrid stand-alone systems.