| Summary: | 碩士 === 東吳大學 === 企業管理學系 === 98 === With the liberalization of financial market,the government relieves all kinds of
financial control continually. Opening new set financial agency set up and loosen the limit
of bank set up branch agency, great amount of bank are established, so the business
competition among banks became fierce, the contend of financial market are just like
having wars. All banks devoted in all their strength in order to take possession of the
market. To expend the market share is their primary goal. The major income of the
general bank in our country is mostly credit business, to create the maximum profit.
However, banks put in their best effort to occupy mortgage market by reducing the
interent rate and expand the market share in order to survive. In the financial agency
consumer loan of our country, mortgage ossupied the largest proportion, in facing the
competition pressure of financial liberalization and internationalization, the bank should
upgrade service quality actively, set up well relationship with the client and then service
various client with exclusive service through client relationship management, in order to
provide the client suitable service, satisfy their demands, raise the contentmentand
loyalism of the client, increase customer and decrease the run off of them, reduce bank
marketing cost, to reach the goal of bank long-term profit and sustainability operate.
The bank should understand and analyze the data of their client; different clients must
have various characteristics and property, group the slient with system. In this research,
we built a group made with effetice segmentation client group through data mining, basic
client information center…etc. Under the group result, using decidion tree, classification
and regression tree and rough sets theory these three methods to compare grouping
accuracy, obtain better grouping method and establish client grouping characteristic and
rule. According to the model prediction, rough sets theory have better grouping result
and clear grouping rule than decision tree and classification and regression tree.
Although decidion tree and classification and regression tree produce rules as well, it
becomes simplify after trimming, therefore, K-means clustering and rough sets theory are
better suggestion for decision makers to use as consumer loans client segmentation
reference. When marketing with new or original client, it is possible to differentiate the
loan product that suits the cliest feature rapidly and increase its desire. It helps the bank
to design loan product for various client, draw up different interest rate and bargain
conditions, pre-assess bank consumer loans at the same time, and reduce risk after loaning,
which creates the maximum profit.
|
|---|
