Ye et al. proposed a concurrent activity recognition technique by analyzing real-time input sensor events to determine their semantic dissimilarity to segment a continuous sensor sequence into fragments, which a fragment corresponds to one ongoing activity [1]. Sensor events is defined as a function with three parameters, each of which refers to reported time of sensor event, location of the sensor and object associated with the sensor. The technique quantify conceptual similarity and time similarity to establish semantic similarity. Conceptual similarity function is built on the least common subsumer (LCS) algorithm which is proposed by Wu et al. [2]. Time similarity function measures the temporal similarity between the sensor events.
[1] Juan Ye, Graeme Stevenson, Simon Dobson, KCAR: A knowledge-driven approach for concurrent activity recognition, Pervasive and Mobile Computing, Available online 22 February 2014, ISSN 1574-1192, http://dx.doi.org/10.1016/j.pmcj.2014.02.003.
[2] Zhibiao Wu and Martha Palmer. 1994. Verbs semantics and lexical selection. In Proceedings of the 32nd annual meeting on Association for Computational Linguistics (ACL ’94). Association for Computational Linguistics, Stroudsburg, PA, USA, 133-138. DOI=10.3115/981732.981751 http://dx.doi.org/10.3115/981732.981751