Research on the analytics of traffic pumping in telecommunications via data science using rehabilitated frog leaping algorithm

Abstract

One kind of telecom crime known as "traffic pumping" occurs when local phone companies artificially overstate the volume of calls flowing into their systems so that they may charge the calling party a greater access fee than their own. Lacking labels for training set makes it difficult to determine whether congestion pumps has occurred. In this study, we suggested a decision-support system based on cluster analysis and decision trees for identifying fraudulent cases. In this study, we use the IBM Telco and Cell2cell datasets. The gathered information can be preprocessed using normalization. When we have collected enough data, we use the rehabilitation frog leaping algorithm (RFLA) to divide up the possible incidents of fraud into distinct groups. Next, we used the cluster participation labels to build a decision tree that led us to the criteria that must be met in order to pursue legal action against the circumstances that raised red flags. Professionals in the field of telecommunications (TC) verify these guidelines in an effort to find a legal remedy against accused offenders. The results are demonstrated and proved the efficiency of the proposed system by comparing it with the conventional techniques.