A Research Review of Adaptive Universal Time Series Forecasting for Intelligent Operation and Maintenance

Amidst the challenges posed by the era of big data, time series forecasting stands out as a pivotal tool for data analysis and management, garnering widespread interest in both its research and application. Methodologies of time series forecasting are reviewed in this dissertation, with a particular focus on advancements in applications based on deep learning and traditional statistical methods, and the adaptive selection algorithms based on operation and maintenance data are discussed. In this research, the key insights from current methodologies are distilled, and recommendations for future research directions and outlook are given, all with the aim for further development of time series forecasting technology.