Mobile Information Systems / 2016 / Article / Tab 10

Research Article

A Mobile Application for Easy Design and Testing of Algorithms to Monitor Physical Activity in the Workplace

Table 10

Small displacement subclassifier performances.
(a) Performance by cumulative indexes

Correctly classified instances 86.65%
Incorrectly classified instances 13.35%
Kappa statistic 0.719
Mean absolute error 0.116
Root mean squared error 0.269
Relative absolute error 35.91%
Root relative squared error 66.92%

(b) Confusion matrix

Classified as
WalkingUpstairsDownstairs

Walking 2787 65 58
Upstairs 105 515 123
Downstairs 96 123 399

(c) Detailed accuracy by class

Precision Recall-scoreClass

0.933 0.958 0.945 Walking
0.733 0.693 0.712 Upstairs
0.688 0.646 0.666 Downstairs

0.862 0.867 0.864 Weighted average