Understanding Human Behaviors by WiFi Data Mining

1
Understanding Human Behaviors by WiFi Data Mining

• Qygroup leaded by Prof. Qiang Yang
• Department of Computer Science and
  EngineeringHong Kong University of Science and
  Technology

Project webpage http//www.cse.ust.hk/vincentz/V
track.html
2
Our Objective

• WiFi Data Mining
• WiFi data are extensively available
• sMobleNet, PCCW, etc..
• Data mining to extract useful patterns
• Clustering, classification, etc.
• Understanding human behaviors
• Users locations (where is he/she?)
• Task 1 Location estimation
• Users activities (what is he/she doing?)
• Task 2 Activity recognition

3
Our Work

• Task 1 Location Estimation
• WiFi signals ? Location
• Semi-supervised learning
• Labeled data insufficiency
• Transfer learning
• Sensor data sharing vary time, devices, etc.
• Task 2 Activity Recognition
• WiFi signals ? activity
• Having class, Having seminar, etc.
• Graphical model
• Sequential information
• Multiple goals in real-world human behaviors

4
Call for Participation

• All of these are based on the collected ltsignal,
  location, activitygt data
• So we will really appreciate it if you can kindly
  help in the data collection!
• Your efforts will greatly contribute to the
  research development of ubiquitous computing!