A decision support system to optimize debt collection assignments

The technological developments let people use mobile phones and benefit from mobile phones in many areas of their lives. People benefit from various services of the operator companies. Therefore, operator companies have an extensive customer base. Yet, collecting the fees of their services from customers can be hard. When the customers regret or delay the payments the operator companies, which serve to millions of customers, face difficulties in legal procedures. The operator companies usually make agreements with the law firms to convey the lawsuits. In this study, a leading GSM operator company wants to know the possibility of finalizing the cases and take prevention on it, when transferring the case files to the law firms. Naive Bayes classifier, decision tree algorithms, k nearest neighbor method, support vector machines, random forest algorithm, and artificial neural network algorithms are examined, and Naive Bayes classification algorithm is used to define the collection difficulty level for the files.