Handbook of Mobility Data Mining, Volume 2: Mobility Analytics and Prediction

Front Cover
Handbook of Mobility Data Mining
Handbook of Mobility: Data Mining Mobility Analytics and Prediction
Copyright
Contents
List of contributors
Preface
Acknowledgments
One - Multi-data-based travel behavior analysis and prediction
1. Introduction
2. Description of mobility big data and travel behavior
2.1 Mobility data mining methods based on heterogenyeousmeans
2.1.1 Based on Bluetooth, WiFi, video detection
2.1.2 Based on GPS and POI
2.2 Definition and description of travel behavior
3. Travel behavior analysis based on mobility big data
3.1 Mobility data processing
3.1.1 Trajectory noise data processing
3.1.2 Analysis of stay point detection
3.1.3 Map matching
3.2 Analysis and prediction of travel behavior
3.2.1 Machine learning applications in activity-travel behavior research
References
Two - Mining individual significant places from historical trajectory data
1. Background
2. Related work
3. Methodology
3.1 Stay location extraction
3.2 Spatial clustering of stay location
3.3 Identify the semanteme of the significant places
4. Application
4.1 Preliminary setting
4.2 Case one: analysis of life pattern changes in the Great Tokyo area
4.3 Case two: analysis of population changes after the fukushima earthquake
4.4 Case three: analysis of the residential location of the park visitors in Tokyo and the surrounding area
References
Three - Mobility pattern clustering with big human mobility data
1. Introduction
2. Related works
3. Methods
3.1 Metagraph-based clustering method
3.1.1 Support graph and topology-attribute matrix construction
3.1.2 Structure constrained NMF and meta-graph space
3.2 Other methods
Preliminary definition
4. Application
4.1 Application case
4.2 Algorithm performances
4.2.1 The computation efficiency
4.2.2 The representational capacity to the mobility pattern differences
References
Four - Change detection of travel behavior: a case study of COVID-19
1. Introduction
1.1 Background
1.2 Related works
1.3 Objectives
2. Methodologies
2.1 Data preprocessing
2.2 Travel behavior pattern change detection
2.3 Data grading
3. Results and analysis
3.1 Individual level
3.2 Metropolitan level
4. Conclusion and discussion
4.1 Summary
4.2 Limitations and future direction
References
Five - User demographic characteristics inference based on big GPS trajectory data
1. Introduction
2. Preliminary
2.1 Definition
2.2 Solving barriers
3. Methodology
3.1 Framework
3.2 Variation inference theory
3.3 Variation inference model construction
3.4 PSO based method (baseline method 1)
3.5 Deep learning-based method (baseline method 2)
4. Case study: experiment in Tokyo, Japan
4.1 Data description
4.2 Baseline settings
4.3 Evaluation metrics
4.4 Overall results
4.5 Evaluation by time use survey data
4.6 Evaluation by built environment demographics
5. Conclusion
References
Further reading
Six - Generative model for human mobility
1. Introduction
1.1 Background
1.2 Problem definition
1.3 Research objective
2. Methodology
2.1 Preliminary
2.2 Framework
3. Experiments
3.1 Descriptions of raw data
3.2 Data preprocessing
3.3 Experimental settings
3.4 Results and visualization
4. Conclusion
4.1 Discussion
4.2 Limitations
References
Further Reading
Seven - Retrieval-based human trajectory generation
1. Introduction
1.1 Background
1.2 Research objective
2. Map-matching as postprocessing
2.1 Framework
2.2 Experiments
3. Metrics for assessment
3.1 Results
3.2 Discussion
4. Retrieval-based model
4.1 Preliminary
4.1.1 Bidirectional long-short term memory
5. K-dimensional tree
5.1 Framework
6. Experiments
6.1 Data description
6.2 Baseline methods and metrics
6.3 Results
7. Conclusion
References
Further reading
Eight - Grid-based origin-destination matrix prediction: a deep learning method with vector graph transformation si ...
1. Introduction
2. Origin-destination matrices
3. Methodology
3.1 Deep learning model-based vector graph transformation loss function
3.2 Grid-based origin-destination matrix prediction model
4. Data generation and study area
5. Result and discussion
5.1 Result of deep learning model-based vector graph transformation loss function
5.2 Result of grid-based origin-destination matrix prediction model
6. Conclusion
References
Nine - MetaTraj: meta-learning for cross-scene cross-object trajectory prediction
1. Introduction
2. Related works
2.1 Social interactions for trajectory prediction
2.2 Multimodality of trajectory prediction
2.3 Meta learning on trajectory prediction
3. Problem description
4. MetaTraj
4.1 Overall architecture
4.2 Subtasks and meta-tasks
4.3 MetaTraj training
4.4 Loss function
4.5 Transformed trajectories
5. Experiments
5.1 Quantitative evaluation
5.2 Ablation studies
5.3 Qualitative evaluation
6. Conclusion
References
Ten - Social-DPF: socially acceptable distribution prediction of futures
1. Introduction
2. Related works
2.1 Social compliant trajectory prediction
2.1.1 Spatiotemporal graphs for trajectory prediction
2.1.2 Multimodal trajectory prediction
2.1.3 Loss functions for trajectory prediction
3. Problem formulation
4. Methodology
4.1 Overall architecture
4.2 Social memory
4.3 Path forecasting
4.4 Loss function
5. Experiments
5.1 Quantitative evaluation
5.2 Qualitative evaluation
6. Conclusion
References
Index
A
B
C
D
E
F
G
H
K
L
M
N
O
P
Q
R
S
T
U
V
W
Back Cover