Modeling and Discovering Occupancy Patterns in Sensor Networks Using Latent Dirichlet Allocation - Foundations on Natural and Artificial Computation - page 487

Table 1. Some obtained topics and the 5 most probable words for each topic (with
k = 90)

Topic 7 Topic 10 Topic 15 Topic 24 Topic 52

words p(w | topic) words p(w | topic) words p(w | topic) words p(w | topic) words p(w | topic)

III6070 0.211401 III6103 0.178530 III6069 0.199488 III6008 0.219808 III6009 0.210557

OOO8070 0.131730 OOO8103 0.132126 OOO8069 0.128993 OOO8008 0.118008 OOO8009 0.123193

IIO7070 0.022237 IOO7103 0.021666 OII3069 0.014108 IIO7008 0.013609 IOO7009 0.011962

OII3070 0.013220 IIO7103 0.021666 IOO7069 0.012204 III8008 0.005343 IIO7009 0.011367

IOO7070 0.013220 III7103 0.013787 IIO7069 0.011843 IOI7008 0.003309 IIO3009 0.004305

Topic 53 Topic 74 Topic 76 Topic 36 Topic 11

words p(w | topic) words p(w | topic) words p(w | topic) words p(w | topic) words p(w | topic)

III6096 0.204948 III6007 0.213525 III6011 0.209088 OOO8007 0.231541 III4023 0.176400

OOO8096 0.131598 OOO8007 0.101672 OOO8011 0.107636 OOO2054 0.193013 OOO7023 0.154432

IIO7096 0.016975 IIO7007 0.013434 III7011 0.042809 III5069 0.017457 OOO8023 0.148823

OII3096 0.015717 III8007 0.007506 OOI8011 0.012842 OOO5102 0.016907 IOO6023 0.030838

IOO7096 0.015717 IIO8007 0.006711 III8011 0.004283 III6069 0.012438 IOO7023 0.027737

Fig. 3. Big picture of occupancy frequencies for all the rooms over the whole data. Each
cell corresponds to a 10 minute interval. Rows indicate specific rooms and columns the
time evolution.

aprob > (0.3)) meaning that this rooms are occupied between 14:00 and 17:00
and empty between 19:00 and 21:00. The topic 36 indicates that when room 7 is
empty from 19:00 to 21:00, room 54 is also empty from 6:00 to 7:00. Finally, the
topic 11 shows a correlation when the room 23 is occupied from 9:00 to 11:00
and empty from 17:00 to 21:00 (OOO7023 and OOO8023).

6 Conclusions and Future Work

The proposed method for modeling the occupancy behaviour of an oce building
is based on the “bag of words” assumption, so each word in a document follows
the ex-changeability assumption. The ex-changeability assumption means that
the words are conditionally independent and identically distributed with respect
to a latent parameter (topics).
A probabilistic model is enough powerful to handle the uncertainty of the
underlying data. The main advantage of the LDA model over other techniques,
is that it allows to obtain a probability distribution of several words in each
topic, following an unsupervised scheme. The obtained results show that the
LDA model is enough powerful to handle the detection of complex patterns over