Estimation of annual average daily traffic volumes using Neural Networks
File, 2013-064-001-019
- Part of
- Algoma University Archives > Algoma University Theses collection > Computer Science series > Estimation of annual average daily traffic volumes using Neural Networks
- Creator
- Mario Adamo
- Date
- 1994
- Physical Description
- 2.23 MB of textual records (PDF)
- General Material Designation
- Electronic record, Textual record
- Language(s)
- English
- Bibliographic Information
- Sault Ste. Marie, Ont.:, OSTMA-COSC-Adamo-Mario-19940402
- Descriptive Notes
- Audience: Undergraduate. -- Dissertation: Thesis (B. A.). -- Algoma University, 1994. -- Submitted in partial fulfillment of course requirements for COSC 4235. -- Includes figures, tables and Graphs. -- Contents: Thesis.
- This study compared the estimations of annual average daily traffic (AADT) volumes using the conventional method(Factors), multiple regression analysis, and the neural network approach. All three approaches were compared using three different classification schemes as well as different duration of traffic counts. The neural network and multiple regression approaches consistently performed better than the conventional approach, and the neural network approach in many cases slightly outperformed the multiple regression approach. Apart from providing a good modeling tool for estimating AADT, the results also provide useful insight into the duration of the short term traffic counts and the classification schemes for the highway sites.