首页--社会科学总论论文--社会学论文--社会结构和社会关系论文--社会心理、社会行为论文

基于社交媒体数据的人类移动模式发现研究

Abstract第4-6页
摘要第7-16页
1 Introduction第16-25页
    1.1 Background第16-17页
    1.2 Current Studies第17-20页
        1.2.1 Displacement Distance第17-19页
        1.2.2 User's Visited Locations第19页
        1.2.3 Regularity,Periodieity and Predictability第19-20页
    1.3 The Development of Data Sources第20-23页
        1.3.1 Survey-based traditional data第21页
        1.3.2 Mobile phone data第21-22页
        1.3.3 Transportation data第22页
        1.3.4 GPS第22-23页
        1.3.5 Location-based social networks第23页
    1.4 The importance of human mobility studies and its application areas第23-24页
    1.5 Chapters and Contributions第24-25页
2 Related Work第25-36页
    2.1 Laws of Human Travel from Bank Notes Dispersal第25-26页
    2.2 Basic Law of Human Mobility Pattern第26-27页
    2.3 Limits of predictability of human mobility第27-28页
    2.4 Universal Model for Mobility Patterns第28-29页
    2.5 Encounters of Familiar Strangers on Transportations第29-31页
    2.6 17 motifs of mobility patterns第31-33页
    2.7 Understanding Human Mobility from Twitter第33-34页
    2.8 Thesis and its substantiation第34-36页
3 Data第36-41页
    3.1 Data Collection第36-38页
    3.2 Data Observation-Collective Analysis第38-41页
        3.2.1 The distribution of the number of tweets of each user第38页
        3.2.2 Weekdays and Weekends第38-39页
        3.2.3 Radius of Gyration第39-41页
4 Methodology第41-48页
    4.1 Assumption第41页
    4.2 Preprocessing Data第41-42页
        4.2.1 Location Identification第41页
        4.2.2 Ruling out improper tweets from the dataset第41-42页
    4.3 Models第42-48页
        4.3.1 κ-Nearest Neighbor第42页
        4.3.2 κ-means第42-45页
        4.3.3 Gaussian Mixture Model(Static)第45页
        4.3.4 Gaussian Mixture Model(Dynamic)第45-46页
        4.3.5 Static and Dynamic Integrated Model(SDIM)第46-48页
5 Experiment第48-67页
    5.1 k-Nearest Neighbors-one location for given time t第48-51页
        5.1.1 Overview for different k第48-49页
        5.1.2 For k=100第49-51页
        5.1.3 Determining k using cross-validation第51页
    5.2 Frequently Visited Location Identification-two locations(home and workplace)第51-60页
        5.2.1 k-means第52-55页
        5.2.2 Home Identification第55-56页
        5.2.3 Gaussian Mixture Model-with two Gaussian components第56-57页
        5.2.4 Gaussian Mixture Model-determining the number of com-ponents第57-60页
    5.3 Location Prediction第60-62页
        5.3.1 Static model-one GMM第60-61页
        5.3.2 Dynamic model-24 GMMs第61页
        5.3.3 Static and Dynamic Integrated Model第61-62页
        5.3.4 Evaluation第62页
    5.4 Entropy Analysis第62-66页
        5.4.1 Entropy with respect to time(Spatial Analysis)第63-64页
        5.4.2 Entropy with respect to Location(Temporal Analysis)第64-66页
    5.5 Applications第66-67页
6 Conclusion and Future work第67-71页
    6.1 Conclusion第67-69页
    6.2 Contribution第69页
    6.3 Future work第69-71页
References第71-75页
Appendix第75-78页
About Author第78-79页
Acknowledgements第79页

论文共79页,点击 下载论文
上一篇:境外非政府组织在华活动的“灰色地带”研究
下一篇:基于聚多巴胺纳米颗粒的肿瘤放化疗协同治疗研究