| Title | The Study on Imputing Method for Incomplete O/D Table |
| Dept | Transportation Technology and Information Division |
| Year | 1997 |
| Month | |
| Price | |
| Summary | The origin-destination (O/D) trip flow is the most basic data for transportation planning and design. As models that estimate O/D matrices from traffic counts are not well established, adopting sampling technique in the O/D trip matrix is still the main approach. The accuracy of O/D data depends on the sampling rate. The higher a sampling rate, is the more precise our data becomes however, it comes with high cost. Consequently, the incomplete O/D table is emerged under the limitation of cost and resource. The purpose of this study is to adopt an imputing method in the incomplete data during survey caused by the sampling process. First, to declare the importance of supplement incomplete O/D matrix, incomplete data were defined. Then methods to investigate O/D trip flow and the sampling technique for different survey were introduced. Because different sampling technique will generate different type of O/D table, the imputing missing values model is explored after classification. The primal analysis shows that certain sampling technique just generate one O/D table, while others produce several O/D tables. For the former, we can determine structure of incomplete data by the pattern of population however, for the latter, their O/D table pattern may be different from the population's. So relative accessibility should be used to examine the existence of structural incomplete data in O/D table. Finally, to increase the flexibility of this model, the imputing missing values models were also introduced according to the type of collected data. |
| Post date | 1997/12/12 |
View count:
120