Generation and quality assessment of route choice sets in public transport networks by means of RP data analysis
Bibliographic record
Abstract
This paper will describe how modeling route choice presumes the generation of a choice set of alternatives that are perceived as available and then are chosen from by travelers. Recent research has posed increasing attention toward the importance of size and composition of choice sets in route choice modeling, and has shown that the generation technique has great impact on route choice model estimates and predictions. This paper has limited knowledge concerning the actual route choices of passengers in public transport networks and the evaluation of the quality of generated choice sets with respect to real life choices. One of the reasons lies in the difficulty to collect data on actual route choices in public transport networks, since a lot of information has to be provided to describe the routes actually used by travelers. In fact, while for private transport it is possible to use global positioning system (GPS) devices to track routes and then map the data to a physical network, for public transport the same method is of little help because relevant information about the lines used is not retrievable with these devices. A problem with GPS is also that signal fall outs in tunnels (metro and sections of the urban rail system). Moreover, GPS devices do not allow obtaining information on the trip purpose, which is another fundamental piece of information for uncovering route choice determinants. This study relies on approximately 2,000 observations of actual route choices in a public transport network, which have been collected by means of a detailed questionnaire gathering all relevant trip information about routes, lines and purposes. The method for choice set generation in a public transport network is based on a timetable probit-based stochastic transit assignment model based on MSA. The method defines a doubly stochastic function that accounts for heterogeneity in both perceived costs and individual preferences. Moreover, the method considers similarities across alternatives and differences in feeder modes. The complexity of the route choice of public transport passengers is therefore represented with a high level of detail. The generated choice sets are tested in terms of quality and number of attractive routes by comparing them with the observed choices. Considering each origin and destination (OD)-pair, attractive routes are defined theoretically as alternatives that travelers would consider, and operatively as the set of alternatives actually selected by all travelers sharing that specific OD-pair. Furthermore, the choice probabilities in the generated choice sets are compared to the actual choice probabilities from the route choice observations. For assessment of the quality of the choice sets, data from the Danish Travel Behavior Survey are used. This survey collects detailed information about route choices of public transport passengers in the Greater Copenhagen area. The dense public network includes trains (regional, urban, local rail), metro and buses (high class and regular). Most travelers have many possible alternative routes, as a result of the combinatorial nature of the problem of combining different modes and different lines. Because of this combinatorial problem, in this dense public network the number of alternative routes can be very high, even though not all possible routes are relevant and attractive to travelers.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.021 | 0.074 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.002 |
| Bibliometrics | 0.005 | 0.004 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.002 | 0.002 |
| Open science | 0.002 | 0.002 |
| Research integrity | 0.001 | 0.002 |
| Insufficient payload (model declined to judge) | 0.003 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".