STATISTICAL METHODS FOR LARGE-SCALE TRANSPORTATION NETWORK TRAFFIC VOLUME FORECASTING

Forecasting is the procedure of making declarations about future events whose actual outcomes have not yet been observed. A lot of decisions are made based on predictions of future unknown events. Knowing the essence of forecasting, it is not hard to interpret what traffic volume forecasting is. Traffic volume forecasting is the process of estimating the number of vehicles that will be on a planned highway in the future. It plays important roles in different aspects of transportation and related field, such as highway level of service analysis, measure of effectiveness, highway improvement and expansion, geometric design and air quality analysis, etc. A good forecast is needed for decision making in future land use and transportation planning. City and county planners can provide useful information about land use planning and projected developments. County engineers may provide information about future county projects that may cause detours and changes in traffic patterns along a trunk highway. Highway designers need forecasted traffic volumes to ensure proper geometric designs. Since short term forecasting has been a hot topic on research, many statistical methods have been used, such as the mean, historical moving average, exponential smoothing, and autoregressive integrated moving average. Among them, Box Jenkins method (Autoregressive Integrated Moving Average) has been found as the best model on forecasting of time series data with seasonality and trend.M.S. in Civil Engineering, December 201