Integrated Moving Average

Integrated Moving Average (IMA) models, a class of time series analysis techniques, aim to forecast future values by incorporating both autoregressive (past values) and moving average (past forecast errors) components. Current research emphasizes the application of IMA models (often in variations like ARIMA, SARIMA, and ARMAX) across diverse fields, including finance, healthcare, and transportation, with a growing focus on comparing their performance against advanced deep learning architectures like LSTMs and Transformers. The effectiveness of IMA models, particularly in conjunction with other techniques like Bayesian regression and model stacking, is being rigorously evaluated to improve prediction accuracy and uncertainty quantification in various real-world applications.