Extending Time Series Forecasting Methods Using Functional Principal Components Analysis.

摘要

Traffic volume forecasts are used by many transportation analysis and management systems to better characterize and react to fluctuating traffic patterns. Most current forecasting methods do not take advantage of the underlying functional characteristics of the time series to make predictions. This paper presents a methodology that uses Functional Principal Components Analysis (FPCA) to create smooth and differentiable daily traffic forecasts. The methodology is validated with a data set of 1,813 days of 15 minute aggregated traffic volume time series. Both the FPCA based forecasts and the associated prediction intervals outperform traditional Seasonal Autoregressive Integrated Moving Average (SARIMA) based methods.