ABSTRACT
Department of Mathematics and Computer Science, University of Southern Denmark
Zhanqing Wu, XianPing Tao
State Key Laboratory for Novel Software Technology, Department of Computer Science and Technology, Nanjing University
Hung Keng Pung
School of Computing, National University of Singapore, Singapore
Jian Lu
State Key Laboratory for Novel Software Technology, Department of Computer Science and Technology, Nanjing University
22.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 318 22.2 Data Preprocessing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 318 22.3 Mining Emerging Patterns For Activity Recognition . . . . . . . . . . . . 319
22.3.1 Problem Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 319 22.3.2 Mining Emerging Patterns from Sequential Activity
Instances . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 319 22.4 The epSICAR Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 320
22.4.1 Score Function for Sequential Activity . . . . . . . . . . . . . . . . . . 320 22.4.1.1 EP Score . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 320 22.4.1.2 Coverage Score . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 321 22.4.1.3 Correlation Score . . . . . . . . . . . . . . . . . . . . . . . . . . 322
22.4.2 Score Function for Interleaved and Concurrent Activities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 322
22.4.3 The epSICAR Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 323 22.5 Empirical Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 324
22.5.1 Trace Collection and Evaluation Methodology . . . . . . . . . . 324 22.5.2 Experiment 1: Accuracy Performance . . . . . . . . . . . . . . . . . . . 325 22.5.3 Experiment 2: Model Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . 326
22.6 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 327
With the rapid advances of sensors and wireless networks, recognizing human activity based on sensor readings has been recently attracting much research interest in the pervasive computing community. A typical application is monitoring activities for the elderly and cognitively impaired people, and providing them with proactive assistance.
