Temporal Regularized Matrix Factorization for High-Dimensional Time Series Forecasting

Time series forecasting plays a critical role in numerous domains, including finance, economics, climatology, and retail. The ability to predict future values based on historical patterns enables better decision-making, resource allocation, and risk management. Traditional approaches to time series forecasting include statistical methods such as autoregressive integrated moving average (ARIMA) models, exponential smoothing, and vector autoregression (VAR)[1]