modelling human mobility, activities, and behaviours for ......query-browse graph for contextual...
Post on 02-Jan-2021
7 Views
Preview:
TRANSCRIPT
Smarter, Resilient, and Fairer Cities
with Deep and Rich Models of Human
Activities and Mobility Behaviours
Flora Salim
Deputy Director, Centre for Information Discovery and Data Analytics
Associate Prof, CS&IT, School of Science, RMIT University, Melbourne, Australia
Humboldt Fellow, University of Kassel, Kassel, Germany
From Human Activity and Behavioural Patternsto Prediction and Optimisation of Urban Resources
Sadri, A., Salim, F.D., Ren, Y., Shao, W., Krumm, J., Mascolo, C. (2018) What Will You Do for the Rest of the Day?
An Approach to Continuous Trajectory Prediction, ACM IMWUT / Ubicomp), Vol. 2, No. 4, 186, Dec 2018
Observations: Morning and afternoon similarity
What Will You Do for the Rest of the Day? An Approach to Continuous Trajectory Prediction
Accurate energy use prediction with deep learning
Hourly average prediction error Weekly average prediction error Monthly average prediction error
H. Song, A. K. Qin & F. D. Salim, Evolutionary Model Construction for Electricity
Consumption Prediction, Neural Computing and Applications, pp. 1-19, 2019.
Pedestrian Traffic Forecasting SystemRMIT & City of Melbourne
Melbourne City Council Pedestrian Counting System
Models tested with datasets from New York JFK, La Guardia, and Newark airports
Outcomes published in:• IEEE Access• K-CAP• PACIS
From Human Mobility to Transport Demand Prediction
using taxi, flight, border control, and weather data
Mornington Peninsula Smart Parking and
Amenities for High Demand Areas
• Visitor demand prediction
• Parking recommendation
• Day trip planner
Parking Availability Prediction with Contextual Features
Predicting Parking
Occupancy in New Urban Areas with Clustering
of Contextual Features, IEEE Transactions on
Intelligent Transportation Systems, under review.
Clustering Big Spatiotemporal-Interval Data
case study: balancing parking demands across the whole city
W. Shao, F. Salim, A. Song, and A. Bouguettaya, “Clustering big spatiotemporal-interval data,” IEEE Transactions on Big Data, 2016.
From Human Mobility to Crime PredictionExample cities: Brisbane, New York, Chicago
Individual Risk Factor Analysis of Visitors
S. K. Rumi, K. Deng, and F. D. Salim, “Theft prediction with individual risk factor of
visitors,” in Proceedings of the 26th ACM SIGSPATIAL International Conference on
Advances in Geographic Information Systems. ACM, 2018, pp. 552–555.
Rumi, S.K., Deng, K., Salim, F.D. (2018), Crime event
prediction with dynamic features, EPJ Data Science
(2018) 7: 43.
Predicting when & where crime will strike next
Multiple Traveling Officer Problem
Shao, W., Salim, F. D., Gu, T., Dinh, T., Chan, J., “Travelling Officer Problem: Managing Car Parking Violations Efficiently Using
Sensor Data”, IEEE Internet of Things Journal, 2018.
Kyle K. Qin, Wei Shao, Yongli Ren, Y., Jeffrey Chan and Flora D. Salim, 2019. Solving Multiple Travelling Officers Problem with
Population-based Optimization Algorithms. Neural Computing and Applications, pp.1-27.
PP
P
P
P
P
Car Leaving
Probability
Model
Changing
State of
Map
Cooperation
P PP
P
Thousands
of parking
bays
P P
Empowering Users with Personalised
Recommendation
© Westfield
From Individual and Group Activities to Personalised
Recommendations Indoors
, Y., Tomko, M., Salim, F.D., Ong, K., Sanderson, M. (2015). “Analyzing Web Behavior in Indoor Retail
Spaces. Journal of the Association for Information Science and Technology (JASIST). Vol. 68, 1, Jul 2015.
Cyber Physical Social (CPS) Contexts → Behaviors
, Y., Tomko, M., Salim, F.D., Ong, K., Sanderson, M. (2015). “Analyzing Web Behavior in Indoor Retail
Spaces. Journal of the Association for Information Science and Technology (JASIST). Vol. 68, 1, Jul 2015.
Analysing User Demographics
with CPS Behaviours
Physical: frequency,
weekdays, duration,
interests in shop
categories
Cyber: WiFi
frequency, search
frequency, what to
browse/search
Social: single, with kids, with another adult, in a group
–age: 18-24, 25-39, 40-54, 55+
–education level: Secondary/high school,
Honours degree?
–income: 0-$18,200, $18,201-$37,000,
$37,001-$80,000, $80,000+
–parental status: having kids?
–shopper category: Inner or Rest of Sydney
resident, CBD Worker, Domestic tourist,
International tourist
Ren, Y., Tomko, M., Salim, F.D., Chan, J., Sanderson, M.. “Understanding the Predictability of User
Demographics from Cyber-Physical-Social Behaviours in Indoor Retail Spaces”. EPJ Data Science 7(1), 2018.
Location-Query-Browse Graph for Contextual Recommendation
Y. Ren, M. Tomko, F. Salim, J. Chan, C. L.A. Clarke and M. Sanderson. “A Location-
Query-Browse Graph for Contextual Recommendation”. IEEE Transactions on
Knowledge and Data Engineering (TKDE). 30(2): 204-218 (2018)
From Cyber-
Physical-Social
Behaviours to Task
and Productivity
Assistance
RMIT-Microsoft Cortana
Intelligence InstituteOverview:
Cortana Intelligence Institute is driving the next-generation of
capabilities for Microsoft’s digital assistant, Cortana. Focused on
researching work-related tasks and using sensors in mobile
phones, the CII team builds a complex multidimensional data set,
used to model and predict user’s work-related tasks.
Impact:
● Task intelligence, to support complex tasks such as tracking
a person's progress on a task, reminders, or assisting with
completion of a task.
● Create a virtual assistant that can manage a calendar,
understand the user, be aware of context, and support
multi-turn dialogues.
Automated Decision-Making
and Society
https://www.arc.gov.au/2020-arc-centre-excellence-automated-decision-making-and-society
Context Recognition and Urban Intelligence (CRUISE), a part
of Centre for Information Discovery and Data Analytics (CIDDA)
Acknowledgment: Our Collaborators
top related