Dergiler / İstatistikçiler Dergisi:İstatistik ve Aktüerya / 2014 / Cilt: 7 - Sayı: 1
Doğrudan Pazarlama Stratejilerinin Belirlenmesinde Veri Madenciliği Yöntemlerinin Kullanımı
- Sayfa
- 1–13
- DOI
- —
Özet
Doğrudan pazarlama, ürünlerin olası müsterilerinin belirlenmesi ve belirlenen müsteri kitlesine bu ürünlerin
Abstract
Direct marketing is the process of identifying possible customers of products and promoting these products tothis specified customer mass. Recently, due to the fact that mass marketing campaigns targeting general publicare not successful, firms give more importance to direct marketing campaigns targeting a specific set ofcustomers. Direct marketing methods are more successful escpecially in banking sector where there is morepressure and competition according to other sectors. Data mining methods are used to increase the success ofdirect marketing campaigns by identifying the factors that effect these campaigns. Thus, these methods provideto direct available resorces and to create a reasonable and true set of potential customers. In this study, wefocus on how direct marketing campaigns can be directed in banking sector by using data mining methodssuch as decision trees, logistic regression, Bayesian networks and support vector machines. Also, we examineclass imbalance problem which frequently encountered in the analysis of this kind of data. As a result, SVMlinear, logistic regression and SVM RBF methods were the most successful methods according to the overallaccuracy metric. Moreover, according to the F measure, logistic regression, SVM RBF and CHAID, andaccording to the matthews correlation coefficient, SVM linear, logistic regression and CHAID methods havebeen identified as the most successful methods, respectively.