Adaptive Seasonal Time Series Models for Forecasting Short-Term Traffic Flow

Abstract

Conventionally, most traffic forecasting models have been applied in a static framework in which new observations are not used to update model parameters automatically. The need to perform periodic parameter reestimation at each forecast location is a major disadvantage of such models. From a practical standpoint, the usefulness of any model depends not only on its accuracy but also on its ease of implementation and maintenance. This paper presents an adaptive parameter estimation methodology for univariate traffic condition forecasting through use of three well-known filtering techniques: the Kalman filter, recursive least squares, and least mean squares. Results show that forecasts obtained from recursive adaptive filtering methods are comparable with those from maximum likelihood estimated models. The adaptive methods deliver this performance at a significantly lower computational cost. As recursive, self-tuning predictors, the adaptive filters offer plug-and-play capability ideal for implementation in real-time management and control systems. The investigation presented in this paper also demonstrates the robustness and stability of the seasonal time series model underlying the adaptive filtering techniques.

A novel three-step procedure to forecast the inspection volume

“…It is worth mentioning that it would be interesting to check the performance of a dynamic linear approach in order to overcome the limitations of a static model such as SARIMA. Some authors have found several advantages and improvements with the use of adaptive linear models (Shekhar and Williams, 2008;Guo et al, 2014Guo et al, , 2015.…”

Accurate and Interpretable Bayesian MARS for Traffic Flow Prediction

“…The SARIMA model improves the predictive accuracy via drawing the periodicity of the traffic data. It constructs the independent variables using the traffic data in the past several intervals together with the historical data in the same intervals in the last week [7], [19]. SARIMA not only provides the short-term causality but also the long-term change rule of traffic state.…”

“…3) SARIMA: The SARIMA model is one of the stateof-the-art parametric techniques and has been successfully applied to the traffic prediction [7], [19], [27]. Through capturing the evident repeating pattern week by week of traffic flow data, the SARIMA introduces weekly dependence relations to the standard ARIMA model and improves the predictive accuracy.…”

“…To estimate the parameters of the SARIMA model, the model is first represented in state-space form. Next, the parameters are updated using adaptive filtering methods [7]. In our implementation, Kalman filter is used because it can achieve the best predictive accuracy according to the research of Lippi et al [19].…”

“…For instance, researchers have taken advantage of temporal historical data to predict short-term traffic flow through Kalman filtering [5], autoregressive integrated moving averaging (ARIMA) [6], seasonal ARIMA (SARIMA) [7], nonparametric regression methods such as the k-nearest neighbor approach [8], and spectral analysis [9]. These methods can be also regarded as univariate methods as they are fed with the univariate historical values for the modeled road.…”