Compact Data Learning For Time-Series Forecasting Systems

This paper focuses on optimizing machine learning-based time-series forecasting models by constructing compact data. Methods for optimizing ML training have been improved and integrated into the development of artificial intelligence (AI) systems. Compact Data Learning (CDL) serves as an alternative practical framework to optimize machine learning (ML) systems by reducing the size of the sampling data. This framework originated from compact data design, providing optimal assets without handling complex big data. Compact Data Learning for Time-Series (CDL-TS) is a novel framework aimed at minimizing forecasting model errors. By utilizing reduced sampling and robust comparison procedures, CDL-TS addresses the challenges of forecasting models on extensive real-time data systems. Additionally' an analytic solution for the M/M/1 queueing system with continuous time parameters under finite time limits is presented in this research. The major contributions are valuable insights and practical techniques for improving model training efficiency, particularly in reducing data volume for continuous time-series data.