Congestion prediction challenge contest held (results)
-Challenging the sophistication of congestion prediction on Expressway using expressway toll and route search data-
- Corporate Top
- Press Room
- Press Release Headquarters
- Congestion prediction challenge contest held (results)
2023年7月26日
東日本高速道路株式会社
東京大学大学院情報学環
NEXCO EAST (Chiyoda-ku, Tokyo) and the Graduate School of Interdisciplinary Information Studies of the University of Tokyo (Bunkyo-ku, Tokyo) have been conducting the "Congestion Prediction Challenge Contest (Announced on January 25, 2023) (hereafter referred to as the “Contest”)”, the final judging and awards ceremony was held on July 6th.
In the contest, there were 163 entries in the modeling division, which competes for model accuracy, and 23 entries in the idea division, which proposes new services using data. Decided.
1. Challenge Contest Results
Winners in the modeling category are:
Accuracy Award
| User ID | Features of the development model | Prediction accuracy * | |
|---|---|---|---|
| First place | yim | A model that performs data expansion to secure the number of training data, creates multiple models, and takes the average value | 0.61235 |
| 2nd place | team_try | A time-series model that predicts the next time period by reflecting holiday patterns and congestion predictions for the previous day and time period | 0.60976 |
| 3rd place | isps737 | A model that focuses on the speed difference due to the presence or absence of traffic congestion and incorporates the average speed difference from the previous day into learning | 0.59816 |
- Prediction accuracy is an index that considers both recall and precision, and takes a value between 0 and 1. The higher the value, the higher the accuracy.
- Recall rate: The percentage of traffic jams that actually occurred that could be predicted to occur
- Precise rate: The ratio of traffic jams that actually occurred to those predicted to occur
modeling award
| User ID | Reason for commendation |
|---|---|
| team_try | From the viewpoint of utilization of route search data, which is the purpose of the contest, we have developed a model that best reflects route search data to improve accuracy. |
審査員コメント（東京大学大学院 情報学環 教授 越塚 登）
In response to biased data, we have developed an excellent prediction model in a short period of time by devoting various ingenuity to algorithms and data processing. Unexpected creative optimization and tuning were carried out in the prediction model presented, and we were able to receive very interesting ideas for improving the accuracy of Expressway traffic congestion prediction.
2. idea department
Winners in the Idea category are:
good idea award
| User ID | idea outline |
|---|---|
| Starmie | For the purpose of avoiding traffic jams waiting for quick chargers, use data such as quick charger usage status and Expressway traffic conditions, etc., and inform the prediction of waiting time for quick chargers to decentralize usage. |
| W5EaSD_2016 | For the purpose of avoiding traffic jams, we will promote excursion tours by providing surrounding sightseeing information using data such as traffic conditions on Expressway and general roads, sightseeing spots and event information in the surrounding area. |
Judge's comment (Noriyoshi Nakanishi, General Manager of ITS Promotion Department NEXCO EAST Management Business Headquarters)
We received a wide range of proposals on topics such as enjoying traffic jams, the environment, SA/PA, and traffic safety. One of the Good Idea Awards is a proposal for responding to electric vehicles, which are becoming more popular. The other is a proposal about measures to promote excursions in collaboration with local communities, such as attractive tourist facilities and cultural facilities, to avoid traffic jams, and is related to the utilization of Expressway.
We will continue to study the effective traffic congestion prediction models and ideas that have been proposed for practical use.
3. judge
Noboru Koshizuka Professor, Interfaculty Initiative in Information Studies, University of Tokyo
Yuya Shibuya, Associate Professor, Center for Spatial Information Science, The University of Tokyo
Mobility Journalist Etsuko Kusuda
NEXCO Research Institute Traffic Environment Research Department Traffic Research Manager Ken Xing Shin Jan
Noriyoshi Nakanishi General Manager, ITS Promotion Department, NEXCO EAST Management Business Headquarters
Four. What is a challenge contest?
This contest is part of the "Joint Research on Data Utilization Efforts" based on the "Research Cooperation Agreement on Information Society Infrastructure" concluded in 2011 by NEXCO EAST and the University of Tokyo Graduate School of Interdisciplinary Information Studies.
One of the things that NEXCO EAST is working on to create a safer and more secure transportation future is traffic congestion prediction. In research on traffic congestion, although the causes and mitigation methods have been found, predictions are made based on many years of experience. This contest is one of the priority projects (6) and (29) of our company's "Aim of the next-generation Expressway that accelerates the realization of an autonomous driving society (concept)" (announced on April 28, 2021). We are soliciting innovative data analysis cases and ideas for the purpose of improving the accuracy of traffic congestion prediction by utilizing big data.
The NEXCO EAST Group has positioned the period from 2021 to 2025 as "a period to contribute to the achievement of the SDGs and transform toward a new future society", and is making various efforts.
We believe that the holding of this Congestion Prediction Challenge Contest will contribute to SDGs goals 3 and 9 as a business activity that will lead to the provision of safe road spaces by alleviating traffic congestion.
To view the PDF file, you need the Adobe Systems plug-in software "Acrobat Reader (Japanese version)". If you do not have Download from here (free)Please use it.