Research Article

Semantic Labeling of User Location Context Based on Phone Usage Features

Algorithm 1

Example: computing features related to phone calls and night stay in different data representations.
Assume that start times and durations are available for
(i) visits: and
(ii) calls: and
(iii) nights: and
visit   (compute for all )
(i) find the smallest index such that
(ii) find the largest index such that
,
 where multiplication with converts the time units to hours
places: home
(i) find set of all visit indices such that  
cumulative sample  : home (compute for all )
Computation of for calling at home:
(i) find set of all visit indices such that
 and
Computation of for night stay at home:
for all  
if exists such that
else if exists such that
end for