Skip to main content
. Author manuscript; available in PMC: 2017 Jun 8.
Published in final edited form as: Trans GIS. 2016 Apr 12;20(6):869–886. doi: 10.1111/tgis.12181

Table 3.

Quality of detected activities of all phone models among different sampling intervals

Sampling
Interval
(seconds)
Number
of
identified
activities
Number
of
reference
activities
Over
detection
rate a
Under
detection
rate a
Mean
location
deviation
(in meters)
Mean
duration
deviation (in
minutes)
Classification
quality index
5 638 448 0.157 0.040 0 3.366 0.7801
10 693 448 0.209 0.096 7.361 5.073 0.4659
15 474 448 0.042 0.058 10.64 3.307 0.9222
20 486 448 0.060 0.063 12.97 3.491 0.8711
25 463 448 0.050 0.074 15.1 3.583 0.86
30 470 448 0.066 0.080 15.69 3.665 0.8248
35 465 448 0.069 0.089 19.61 4.18 0.7615
40 469 448 0.072 0.098 18.02 4.004 0.7673
45 448 448 0.060 0.100 20.12 4.279 0.7554
50 419 448 0.076 0.145 20.19 4.714 0.6734
55 380 448 0.058 0.179 24.21 4.905 0.6363
60 359 448 0.070 0.203 22.28 4.042 0.6707
65 308 448 0.055 0.252 21.45 5.752 0.5419
70 295 448 0.068 0.283 22.2 5.924 0.4907
75 294 448 0.109 0.313 29.38 5.613 0.4165
80 353 448 0.099 0.283 23.38 3.926 0.5729
85 316 448 0.104 0.319 32.92 5.653 0.3873
90 267 448 0.112 0.379 25.85 3.501 0.5127
95 260 448 0.096 0.384 33.06 4.802 0.4024
100 246 448 0.089 0.404 28.73 5.192 0.3887
a

Over detection rate: calculated by dividing the number of over detection cases by the number of identified activities;

a

Under detection rate: calculated by dividing the number of under detection cases by the number of baseline activities.