应用简易支持向量机(SSVM)进行客户流失预测,以提高机器学习方法的预测能力。以国外电信公司客户流失预测为实例,与最近邻算法(NPA)进行了对比,发现该方法在获得与NPA近似准确率的条件下,所花费的时间和时间增加值远小于NPA,是研究客户流失预测问题的有效方法。
To improve the prediction abilities of machine learning methods, this paper applied a simple support vector machine(SSVM) to customer churn prediction. The method was compared with NPA regarding customer churn prediction for foreign telecommunication carrier. It was found that the method need less time and adding time with the consistent precision, and provided an effective measurement for customer churn prediction.