ZEMCH 2015 - International Conference Proceedings | Page 252

variations. However, none of them indicate any significant change except in the case of House C,
in which SVM boosted its performance from around 40% for 30sec timeslices up to over 60% for
the bigger timeslices of 10 minutes. This is also an indication of the adaptability of the support
vectors to different input spaces.
Figures 5 to 7 show these variations for each of the scenarios included in this dataset; using all
three models (SVM, HMM and kNN).

Figure 5: TA approach with timeslice variations from 30 seconds to 10 minutes. Dataset1 House A.

Figure 6: TA approach with timeslice variations from 30 seconds to 10 minutes.
Dataset1 House B.

250

ZEMCH 2015 | International Conference | Bari - Lecce, Italy