The current number of mobile subscribers reached 50 million people can be seen as a nearly saturated state. To attract new customers means that coming to take away from the competition’s customers, so there are the fierce competition to attract to customers. So if there are signs of churn, it should be quickly analysed to avoid moving to competition. This paper estimated the churn of mobile users by using Competing risks model and Bayes’ theorem, and proposed how to prevent it. Proposed method was verified through the experiment, and it has been shown that the results of using the model is effectively predict the probability of churn of mobile users.

The work on the eutropication of Youngsan river was conducted from Jul. 1997 to Jun. 1998. The value of water temperature in the study area varied from 4℃ to 32℃ with an annual mean of 17.5℃. DO varied from 4.0㎎/ℓ to 13.5㎎/ℓ with an annual mean 10.2 ㎎/ℓ. BOD varied from 0.8㎎/ℓ to 6.4㎎/ℓ with an annual mean 3.1㎎/ℓ. COD varied from 4.6㎎/ℓ to 9.0㎎/ℓ with an annual mean 6.38㎎/ℓ. Nitrate varied from 1.58㎎/ℓ to 6.77㎎/ℓ with an annual mean of 3.475㎎/ℓ. Total-nitrogen varied from 2.14㎎/ℓ to 8.38㎎/ℓ with an annual mean of 5.083㎎/ℓ. Total-phosphate varied from 0.035㎎/ℓ to 0.588㎎/ℓ with an annual mean of 0.128㎎/ℓ. Chl.a varied from 1.8㎍/ℓ to 75.0㎍/ℓ with an annual mean of 19.55㎍/ℓ. The euthrophic state index of Carlson^16) were showed 61.8(TP) and 71(SD) value