Open access
Date
2014-02-28Type
- Report
ETH Bibliography
yes
Altmetrics
Abstract
This deliverable provides an overview over the current status of the work in Work Package 4 “Automated Travel Mode and Trip Purpose Detection”. The deliverable is focused on Task 4.3 Development and implementation of trip purpose imputation. The goal of this task is to implement the detection of trip purposes using temporal and spatial patterns observed in the GPS data. That includes timings and durations of activities and frequencies of visited places. In this deliverable we introduce trip purpose identification using random forests, a machine- learning approach based on decision trees. Analysis is done using GPS and accelerometer data collected by 156 participants, taking part in a one-week travel survey in Switzerland completed in 2012. Results show that random forests provide robust trip purpose classification. For ensemble runs, a share of correct predictions between 80 and 85 % was achieved. Different setups of the classifier are possible, and sometimes required by the application context. The classifier’s training set and its input variables (feature set) are defined in various ways. Four relevant setups are tested here. Show more
Permanent link
https://doi.org/10.3929/ethz-b-000091025Publication status
publishedJournal / series
Peacox – Persuasive Advisor for CO2-reducing cross-modal trip planningVolume
Publisher
Austrian Institute of Technology (AIT)Organisational unit
03521 - Axhausen, Kay W. (emeritus) / Axhausen, Kay W. (emeritus)
02226 - NSL - Netzwerk Stadt und Landschaft / NSL - Network City and Landscape
02655 - Netzwerk Stadt u. Landschaft ARCH u BAUG / Network City and Landscape ARCH and BAUG
Related publications and datasets
Is part of: https://doi.org/10.3929/ethz-b-000086724
More
Show all metadata
ETH Bibliography
yes
Altmetrics