Recent developments in understanding and modelling the behaviour of individuals in a transport context have generated a lot of discussion of the need to identify the markets within which a range of transport policies can be directed. There has been a large amount of research into the use of variables selected as segmentation variables on a priori argument in identifying variations in traveller behaviour, but a limited amount of work into a systematic segmentation for the total population in order to identify appropriate market segments (or groups of individuals) with respect to the particular issue under investigation. Since there are as many market segments as there are issues, it is inappropriate to generalise the role of variables found to be appropriate explanations of variations in behaviour with respect to a particular issue. This paper discusses the general problem of classification as a prior requirement to estimating, and then illustrates the application of a technique designed to identify types of groups of travellers who respond relatively similarly to an issue such as the number of trips to the shops at a particular time of day. The method has the prime aim of generating mutually exclusive sub-groups of the total sample which are individually used in estimation of choice models (each model being one of a set where prior classification has assisted in reducing the amount of variance to be explained by each choice model) and thus improving the predictive capability of the overall set of choice models. Since each group is different, then different policies might be appropriate in achieving a desired outcome. This approach is consistent with developments designed to produce improved aggregate travel demand models which have not lost, prior to estimation, a large amount of the variance to be explained, as is the case with traditional aggregate models. /Author/TRRL/

Media Info

  • Features: Figures; References; Tables;
  • Pagination: p. 23-30
  • Serial:
    • Volume: 8
    • Issue Number: 1

Subject/Index Terms

Filing Info

  • Accession Number: 00164182
  • Record Type: Publication
  • Source Agency: ARRB
  • Files: ITRD, TRIS, ATRI
  • Created Date: Dec 27 1977 12:00AM