Intelligent Signal Analysis Using Case-Based Reasoning for Decision Support in Stress Management - Computational Intelligence in Healthcare: Advanced Methodologies - page 172

the test, hence it has been agreed with the clinician that this step should not been con-
sidered. Classification of individual sensitivity to stress based on “degree of change”
is shown in Fig. 21.
As seen in Fig. 22 each step of the calibration is divided into one minute time in-
tervals (Step 3 is 4 minutes in duration and 4 features are extracted) while each fea-
ture contains 120 sample data (time, temperature).

1

9

10

2

11

12

8

3

4

7

5

6

Fig. 22. Changes in FT data against time during different stress and non-stress conditions

Thus 12 features are extracted from 5 steps (step 2 to 6) and named as Step2_Part1,
Step2_Part2, Step3_Part1 etc. First, a slope of the linear regression line has been
calculated through the data points by using Equation 3 for each extracted feature from
the measurement.

n

∑

( x − x )( y − y )

(3)

slope =

i = 0

f

n

∑

( x − x )

2

i = 0

Where f denotes the nu mber of features (1 to 12 see Fig. 22), i is the number of sam-

x , is the average of the samples. Then the slope value is con-
verted to arctangent as a value of angle in radians ( -pi/2 to +pi/2 ) and finally ex-
pressed arctangent values in degrees by multiplying 180/PI where PI is 3.14 as a
standard value. So these 12 features contain degree values comprising of 120 sample
data (time, temperature). Instead of keeping the sample data, these degree values are
used or represented as features. Five other features which have also been extracted
from the sensor signal are: start temperature and end temperature from step2 to step6,
minimum temperature of step3 and step5, maximum temperature of step4 and step6,
and the difference between ceiling and floor. Finally, 17 (12+5) features are extracted
(Table 4) automatically from the fifteen minutes (1800 samples) of FT sensor signal.
A new case is then formulated from these extracted features (Table 4). Finally,
this new case is passed to the CBR cycle to use it in an automated classification
scheme.

y

ples (1 to 120) and