CAJ | 학술논문

为了在激烈的市场竞争中取胜,电信企业意识到必须将客户分类,针对不同的客户研究相应的营销策略,DBSCAN算法能够实现客户分类,但对初始参数E和MinPts的取值非常敏感,不同的取值将产生不同的聚类结果,通过对DBSCAN算法进行改进,实现了更加准确和全面的客户分类。
위료재격렬적시장경쟁중취성,전신기업의식도필수장객호분류,침대불동적객호연구상응적영소책략,DBSCAN산법능구실현객호분류,단대초시삼수E화MinPts적취치비상민감,불동적취치장산생불동적취류결과,통과대DBSCAN산법진행개진,실현료경가준학화전면적객호분류。
In order to win in the fierce market competition, telecommunications enterprises has realized the importance of customer classification. Different customers accept different marketing strategies. DBSCAN algorithm can achieve customer classification, but the initial parameters E and MinPts values are very sensitive. Different values will produce different clustering results; An improved DBSCAN algorithm is proposed in order to achieve a more accurate and comprehensive customer classification.