Page 83 - Anthropometry, Apparel Sizing and Design

112 Anthropometry, Apparel Sizing and Design

obtained from the tree diagram. The decision tree (refer to Fig. 4.5) classified the clus-
ter groups with correctness of 90% for age 13–17. Hence this confirms that the sample
groups are classified correctly into cluster groups using k-means technique.
For this section the same dependent and independent variables are used to classify
the cluster groups. The dependent variables are the cluster groups for each individual
sample group. The independent variables are the primary control variables (key
dimensions) of each sample group specific to either the upper body or lower body.
The observations are seen in this female sample group for the upper body where
the rule of this tree model is as follows: the maximum depth of the tree extends
two levels beneath the root node at the minimum sample; the parent node is 100,
and the child node is 50; all is the same for sample groups. The same observation
is also seen in the predictive variables where the girth variables are identified as
the predictive variables for female age 13–17.

4.6.1 Decision tree for female (13–17years old) upper body
The results of the tree (Fig. 4.5) show that bust girth is the only predictor for the large
size. The female samples in this group were split into two bust measurements. Those
with bust measurements >86.3cm are most likely to be clustered under the large size.
Those with bust measurement <86.3cm and >161.4cm tall are likely to be grouped as
the medium size. Those <154cm tall with bust <86.2cm are likely to be grouped
under small size. The profile of the clusters is shown in Table 4.8.
This tree verified that the cluster groups were correctly segmented using the cluster
analysis method. The next step is to develop sizing system based on these cluster
groups.

4.7 Body size determination for sizing system

In this study the sizing system was designed according to the design limit that accom-
modated 90% of the population. Using design limit means taking into consideration
the values that might have an impact on the total reliability and practicality of clothing
production for manufacturers (Abtew et al., 2017). Designing according to limit also
prevents sizes to be produced unnecessarily and impractical use for only few people.
The sizing system is built according to the percentile values of 5th and 95th.

4.7.1 Female sizing system 13–17years old for upper body
garments

Fig. 4.6 showed the distribution of sizes according to the body types for female aged
between 13 and 17years old; the control variables for upper body are height and bust
girth. The size interval for height is 8cm and for bust and girth is 6cm. The range for
height in this population is 142–166cm, which is the same for the upper and lower
body. The range is divided into five subgroups using an interval of 6cm, which is
142, 148, 154, 160, and 166cm. For the bust the range is from 70 to 98cm; this range