Xiaolei Li, Zhenhui Li, Jiawei Han, Jae-Gil Lee. 1. Motivation 2. Anomaly Definitions 3. Algorithm 4. Experiments 5. Conclusion.

Slides:



Advertisements
Similar presentations
15th CTI Workshop, July 26, Smart Itinerary Recommendation based on User-Generated GPS Trajectories Hyoseok Yoon 1, Y. Zheng 2, X. Xie 2 and W.
Advertisements

Entry task 1) What can be concluded from the following pattern? = 15 = 3 × = 20 = 4 × = 25 =
Speed vs. Velocity.
Incremental Clustering for Trajectories
Finding Optimal Solution of 15 puzzle B NTUCSIE Kai-Yung Chiang.
© 2007 Cisco Systems, Inc. All rights reserved.Cisco Public ITE PC v4.0 Chapter 1 1 Distance Vector Routing Protocols Routing Protocols and Concepts –
Outline Minimum Spanning Tree Maximal Flow Algorithm LP formulation 1.
ADAPTIVE FASTEST PATH COMPUTATION ON A ROAD NETWORK: A TRAFFIC MINING APPROACH Hector Gonzalez, Jiawei Han, Xiaolei Li, Margaret Myslinska, John Paul Sondag.
An Interactive-Voting Based Map Matching Algorithm
Driving with Knowledge from the Physical World Jing Yuan, Yu Zheng Microsoft Research Asia.
Finding the Sites with Best Accessibilities to Amenities Qianlu Lin, Chuan Xiao, Muhammad Aamir Cheema and Wei Wang University of New South Wales, Australia.
Urban Computing with Taxicabs
Ziming Zhang, Yucheng Zhao and Yiwen Wan.  Introduction&Motivation  Problem Statement  Paper Summeries  Discussion and Conclusions.
Learning Trajectory Patterns by Clustering: Comparative Evaluation Group D.
On Map-Matching Vehicle Tracking Data
Delay bounded Routing in Vehicular Ad-hoc Networks Antonios Skordylis Niki Trigoni MobiHoc 2008 Slides by Alex Papadimitriou.
Constructing Popular Routes from Uncertain Trajectories Ling-Yin Wei 1, Yu Zheng 2, Wen-Chih Peng 1 1 National Chiao Tung University, Taiwan 2 Microsoft.
Adaptive Fastest Path Computation on a Road Network : A Traffic Mining Approach Hector Gonzalez Jiawei Han Xiaolei Li Margaret Myslinska John Paul Sondag.
Critical Analysis Presentation: T-Drive: Driving Directions based on Taxi Trajectories Authors of Paper: Jing Yuan, Yu Zheng, Chengyang Zhang, Weilei Xie,
Motion Detection And Analysis Michael Knowles Tuesday 13 th January 2004.
Trajectories Simplification Method for Location-Based Social Networking Services Presenter: Yu Zheng on behalf of Yukun Cheng, Kai Jiang, Xing Xie Microsoft.
Tracking Moving Objects in Anonymized Trajectories Nikolay Vyahhi 1, Spiridon Bakiras 2, Panos Kalnis 3, and Gabriel Ghinita 3 1 St. Petersburg State University.
Trip Planning Queries F. Li, D. Cheng, M. Hadjieleftheriou, G. Kollios, S.-H. Teng Boston University.
1 Vehicular Sensor Networks for Traffic Monitoring In proceedings of 17th International Conference on Computer Communications and Networks (ICCCN 2008)
Algorithm: For all e E t, define X e = {w e if e G t, 1 - w e otherwise}. Measure likelihood of substructure S by. Flag S as anomalous if, where is an.
1 Indexing Large Trajectory Data Sets With SETI V.Prasad Chakka Adam C.Everspaugh Jignesh M.Patel University of Michigan Presented by Guangyue Jia.
Reducing Uncertainty of Low-sampling-rate Trajectories Kai Zheng, Yu Zheng, Xing Xie, Xiaofang Zhou University of Queensland & Microsoft Research Asia.
Tokyo Research Laboratory © Copyright IBM Corporation 2009 | 2009/04/03 | SDM 09 / Travel-Time Prediction Travel-Time Prediction using Gaussian Process.
Sensys 2009 Speaker:Lawrence.  Introduction  Overview & Challenges  Algorithm  Travel Time Estimation  Evaluation  Conclusion.
A Geometric Framework for Unsupervised Anomaly Detection: Detecting Intrusions in Unlabeled Data Authors: Eleazar Eskin, Andrew Arnold, Michael Prerau,
Extracting Places and Activities from GPS Traces Using Hierarchical Conditional Random Fields Yong-Joong Kim Dept. of Computer Science Yonsei.
Bei Pan (Penny), University of Southern California
Section 2.2 Echelon Forms Goal: Develop systematic methods for the method of elimination that uses matrices for the solution of linear systems. The methods.
Mapping and analysis for public safety: An Overview.
Tonghong Li, Yuanzhen Li, and Jianxin Liao Department of Computer Science Technical University of Madrid, Spain Beijing University of Posts & Telecommunications.
Energy-Aware Scheduling with Quality of Surveillance Guarantee in Wireless Sensor Networks Jaehoon Jeong, Sarah Sharafkandi and David H.C. Du Dept. of.
1 SUPPORT VECTOR MACHINES İsmail GÜNEŞ. 2 What is SVM? A new generation learning system. A new generation learning system. Based on recent advances in.
Background Subtraction based on Cooccurrence of Image Variations Seki, Wada, Fujiwara & Sumi Presented by: Alon Pakash & Gilad Karni.
A Firewall for Routers: Protecting Against Routing Misbehavior1 June 26, A Firewall for Routers: Protecting Against Routing Misbehavior Jia Wang.
Clustering Moving Objects in Spatial Networks Jidong Chen, Caifeng Lai, Xiaofeng Meng, Renmin University of China Jianliang Xu, and Haibo Hu Hong Kong.
Siyuan Liu *#, Yunhuai Liu *, Lionel M. Ni *# +, Jianping Fan #, Minglu Li + * Hong Kong University of Science and Technology # Shenzhen Institutes of.
AAAI 2011, San Francisco Trajectory Regression on Road Networks Tsuyoshi Idé (IBM Research – Tokyo) Masashi Sugiyama (Tokyo Institute of Technology)
INFERRING HUMAN ACTIVITY FROM GPS TRACKS Sun Simiao.
A P2P-Based Architecture for Secure Software Delivery Using Volunteer Assistance Purvi Shah, Jehan-François Pâris, Jeffrey Morgan and John Schettino IEEE.
Trajectory Outlier Detection: A Partition-and-Detect Framework1 04/08/08 April 8, 2007 Trajectory Outlier Detection: A Partition-and-Detect Framework Jae-Gil.
Crowd sourcing techniques and applications for ITS Limitations and possibilities Dionisis Kehagias Senior Researcher at Information Technologies Institute.
Consensus Extraction from Heterogeneous Detectors to Improve Performance over Network Traffic Anomaly Detection Jing Gao 1, Wei Fan 2, Deepak Turaga 2,
Trajectory Data Mining Dr. Yu Zheng Lead Researcher, Microsoft Research Chair Professor at Shanghai Jiao Tong University Editor-in-Chief of ACM Trans.
Trajectory Data Mining Dr. Yu Zheng Lead Researcher, Microsoft Research Chair Professor at Shanghai Jiao Tong University Editor-in-Chief of ACM Trans.
Trajectory Data Mining
U of Minnesota DIWANS'061 Energy-Aware Scheduling with Quality of Surveillance Guarantee in Wireless Sensor Networks Jaehoon Jeong, Sarah Sharafkandi and.
Anomaly Detection. Network Intrusion Detection Techniques. Ştefan-Iulian Handra Dept. of Computer Science Polytechnic University of Timișoara June 2010.
Example Apply hierarchical clustering with d min to below data where c=3. Nearest neighbor clustering d min d max will form elongated clusters!
Motion Segmentation at Any Speed Shrinivas J. Pundlik Department of Electrical and Computer Engineering, Clemson University, Clemson, SC.
2013 IEEE 14th International Conference on Mobile Data Management Authors: 1. Jiansu Pu 2. Siyuan Liu 3. Ye Ding 4. Huamin Qu 5. Lionel Ni By: Farah Kamw.
2 Overhead & channel use for multiple messages. Unused elements in common message set. Security? A new message for each new application? Where to change.
Presented by: Shahab Helmi Spring 2016
Presented by: Mi Tian, Deepan Sanghavi, Dhaval Dholakia
Slicing and dicing the 153-year record of monthly
Online Conditional Outlier Detection in Nonstationary Time Series
Traffic Sign Recognition Using Discriminative Local Features Andrzej Ruta, Yongmin Li, Xiaohui Liu School of Information Systems, Computing and Mathematics.
Chen Jimena Melisa Parodi Menashe Shalom
Mining Spatio-Temporal Reachable Regions over Massive Trajectory Data
Intra-Domain Routing Jacob Strauss September 14, 2006.
BGP update profiles and the implications for secure BGP update validation processing Geoff Huston PAM April 2007.
Latent Space Model for Road Networks to Predict Time-Varying Traffic
Motion Segmentation at Any Speed
Predicting Traffic Dmitriy Bespalov.
Automatic Segmentation of Data Sequences
Improving Transportation Inventories Summary of February 14th Webinar
Presentation transcript:

Xiaolei Li, Zhenhui Li, Jiawei Han, Jae-Gil Lee

1. Motivation 2. Anomaly Definitions 3. Algorithm 4. Experiments 5. Conclusion

Many moving object trajectories being tracked in road networks E-ZPass system in major cities GPS trackers Finding unusual events or anomalies Useful for city traffic control Homeland security

1. Motivation 2. Anomaly Definitions 3. Algorithm 4. Experiments 5. Conclusion

An observation (or a set of observations) which appears to be inconsistent with the remainder of that set of data. Anomalies: (1) other points in the dataset (2) temporal history of itself (3) temporal history of other points in the dataset.

Unusual street segments Speed of vehicles Load of traffic Association with other features Example: A normal street ?

University Ave. usually has the similar traffic pattern as Lincoln Ave. and Neil St. Similar speed patterns throughout the day On Nov. 30th, average speed on University Ave. drops from 40 to 25. University Ave. is an anomaly on that day? Might be two courses: Traffic accident / road construction slows down traffic Snow! Global effect Detect the similar change on Lincoln Ave. and Neil. St

Compare with historical data of itself Compare with other data using time as a dimension on entire time span

1. Motivation 2. Anomaly Definitions 3. Algorithm 4. Experiments 5. Conclusion

Traffic Data Single Edges Edge Sequences Routes Temporal Vector Construction 1 Temporal Similarity Vectors 23 m … Temporal Anomaly Detection … Input Features And Similarity Fcn

On day I, road Y is a similar neighbor of road X: X and Y have been similar over several days until day I Question: how to maintain similar neighbors incrementally without looking back into history?

Maintain a vector for every edge Vector records similarity between the particular edge and all other edges in the road network Vector is updated at each time step Values in vector indicate historical similarity Drastic change in vector indicates anomaly Time Period 1Time Period 2 Time Period 3 Time Period m …

Time Period 1Time Period 2 Time Period 3 Time Period m … 1) How to update?2) How to detect anomaly?

Time Period 1Time Period 2 Time Period 3 Time Period m … 1) How to update? Intuition: Historical Similarity Day is similarity Updated Similarity High Increase lightly HighLowDecrease heavily LowHighIncrease heavily Low Decrease lightly

Time Similarity Incremental similarity decreases Sharp decrease For dissimilarity dissimilar similar

Measure anomaly by calculating similarity between Temporal Neighbor Similarity vector and daily neighbor vector Rank daily anomalies by dissimilarity amount Report top k anomalies every day

1. Motivation 2. Anomaly Definitions 3. Algorithm 4. Experiments 5. Conclusion

Data: 24 days of moving taxicab data in San Francisco area in July, ,000 separate trips 33 million road segment traversals 100,000 distinct road segments

1. Motivation 2. Anomaly Definitions 3. Algorithm 4. Experiments 5. Conclusion

Contribution: A new temporal outlier detection definition. General outlier detection idea Incremental outlier detection Update similarity vector incrementally Further improvement: How to evaluate outliers? How to solve the sparse data problem?