A tutorial on hidden Markov models and selected applications in speech recognition
Readings in speech recognition
Mining models of human activities from the web
Proceedings of the 13th international conference on World Wide Web
The Google Similarity Distance
IEEE Transactions on Knowledge and Data Engineering
Accurate activity recognition in a home setting
UbiComp '08 Proceedings of the 10th international conference on Ubiquitous computing
Real world activity recognition with multiple goals
UbiComp '08 Proceedings of the 10th international conference on Ubiquitous computing
Location-Based Activity Recognition
KI '07 Proceedings of the 30th annual German conference on Advances in Artificial Intelligence
Unsupervised activity recognition using automatically mined common sense
AAAI'05 Proceedings of the 20th national conference on Artificial intelligence - Volume 1
CIGAR: concurrent and interleaving goal and activity recognition
AAAI'08 Proceedings of the 23rd national conference on Artificial intelligence - Volume 3
Location-based activity recognition using relational Markov networks
IJCAI'05 Proceedings of the 19th international joint conference on Artificial intelligence
An 'object-use fingerprint': the use of electronic sensors for human identification
UbiComp '07 Proceedings of the 9th international conference on Ubiquitous computing
Using a live-in laboratory for ubiquitous computing research
PERVASIVE'06 Proceedings of the 4th international conference on Pervasive Computing
| Hi-index | 0.00 |
The fundamental problem of the existing Activity Recognition (AR) systems is that these require real-world activity data to train the underneath activity classifier. It significantly reduces the applicability and scalability of the system. An AR system trained in an environment would only be applicable to that environment and would not be able to recognize new activities of interest. To overcome such difficulties, in this paper we propose a simple and ubiquitous sensor based AR system that uses web activity data to train its classifier. It would work to almost any environment and would be scalable by its very design. Given a set of activities to monitor, object names with embedded sensors and their corresponding locations, the ARHMAM first mines activity data from web, and uses these to build a Hidden Markov Model (HMM). In comparison with the existing web data based AR systems, it has the following advantages: (1) it uses more strong activity model, (2) it reduces the mining time significantly. It is observed that the class accuracy of activity recognition of our system for a real-world dataset is more than 64%, which is 20% more in comparison with its counterpart. Additionally, the mining time complexity is far better than its counterpart.