An improved algorithm for matching partial registration numbers

In order to identify route choice, registration number surveys were carried out and the numbers were then matched. Because of the nature of the routes being investigated, it was found that the level of spurious matches was large relative to the number of tripmakers. This cast considerable doubt on the estimated number of tripmakers and their mean journey times. Accordingly, it was necessary to devise an algorithm which minimised the effect of such errors. This algorithm and a method of estimating the number of residual spurious matches is discussed in this paper and is compared with the original algorithm for an actual example.