Information Technology Reference
In-Depth Information
Fig. 24. Comparison results of the classification for the five test groups (i.e. TG-A, TG-B, TG-
C, TG-D, TG-E and average of all the groups)
Fig. 24 illustrates a comparison for each group on the basis of the number of cases
and correctly classified cases. In Fig. 24, the blue line represents the total number of
the cases in a group and the red line shows the number of cases correctly classified by
the system for the corresponding group. For each case in each test group, the CBR
system correctly classifies 80%, 86%, 78%, 82%, 86% of the cases. So, from the
experimental result, it can be observed that the classification system correctly classi-
fies cases with an average accuracy of 82%.
7 Discussions and Conclusions
In this chapter, a combined approach based on a calibration phase and CBR to provide
assistance in diagnosing stress is proposed using data analysis from finger tempera-
ture sensor readings. Finger temperature is an indirect measurement of stress and
relaxation and as shown in this chapter. It contains information on changes to a per-
son's stress level which can be measured using a low cost sensor. CBR is a method
which enables the diagnosis of stress despite large variations between individuals.
The calibration phase also helps to establish a number of individual parameters. Fuzzy
logic incorporated into the CBR system for similarity matching between cases reduces
the sharp distinction between cases and provides a more reliable solution.
From the experimental result, it is observed that the classification result on average
is 82% for all the three sets of cases using the same feature sets. Fig. 24 suggests
that using extracted features the CBR system can classify a promising number of
cases and that only a few cases are misclassified out of the total number of the cases.
Search WWH ::




Custom Search