| Summary: | 碩士 === 國立中山大學 === 企業管理研究所 === 82 === The research focus on “fuzzy cluster analysis”that has been
applied effectively in “pattern recognition”, to discuss and
compare the“fuzzy cluster analysis”with the traditional
cluster analysis ( nonoverlapping cluster analysis ) ,and also
applied fuzzy cluster analysis to segment the collegians''
electr- onic dictionary consuming market in Kaohsiung as an
empirical study . After reviewing relative lectures , comparing
the difference between the fuzzy cluster analysis and the
traditional cluster analysis on their theories , the research
applied FCM algorithm fuzzy cluster analysis）and two stages
cluster analysis（non- oveerlapping cluster analysis ）
respectively to segment the col- legians'' electronic dictionary
consuming market in Kaohsiung and compared their market
segmenting results,acquiring the conclu- sions as below : 1.
After comparing the market segmenting results , found that both
results are similar to each other , and the cluster res- ult of
fuzzy cluster analysis has certain cluster reliablity . 2.
Consulting the cluster result of fuzzy cluster analysis , we
can try to evaluate the market segments,and divide the market
to several overlapping segments by the memberships. So the
empirical study of the research has revealed that fuzzy cluster
analysis not only can accomplish the work of non- overlapping
clustering but also through the memberships provide some
messages that traditional cluster analysis can not provide.
Therefore the research serves an example of the feasibility of
fuzzy cluster analysis applied on market segmentation.
|
|---|
