How do RNNs Generalize AR-Models?
What is an AR Model? An Auto Regressive Model, or simply an AR-Model, is a statistical model for time series. The main assumption is that the value in time $t+1$ is a linear combination of the $p$ previous values. The $p$ parameter is called the order of the model. Hence an auto regressive model of order $p$ can be concisily referred to as an $AR(p)$ model. Given $\{x_t\}_{t=0}^{n}$ data points the main assumption of an AR($p$) model is expresed through the following equation: ...