One of the most often expressed concerns in the telecom industry is internationalrnrevenue sharing fraud. Fraudsters are motivated to generate trafficrnfor their services without paying the originating network. Existing detectionrnmethods include monitoring call patterns and blocking high-risk range listsrnobtained from various telecom associations. However, these approaches havernissues that make them ineffective, time-consuming techniques, leading to financialrnlosses before the call is blocked as they are frequently bypassed, renewrnhigh-risk range list numbers repeatedly, and make lists out of date.rnThe goal of this paper is to develop an international revenue sharing fraudrnmodel for real-time detection of missed call fraud generation schemes using anrninternational voice call detail record in an hourly manner to minimize detectionrntime as well as revenue loss. To do so, data is collected, data preprocessingrnis performed, and relevant attributes are determined. As classifiers, supportrnvector machines, artificial neural networks, and random forests are utilized tornclassify the data set for fraud and normal call transactions.rnThe findings indicate that random forest techniques outperform in terms of fmeasure,rnaccuracy, receiver operating characteristic curve, the time required forrnbuilding, and inference time in both training modes (percentage split and 10-rncross-validation). It performed with 97.00% accuracy and also achieved comparablernaccuracy to the state of the art. Consequently, the support vector machinernand neural network multilayer perceptron classification algorithms are foundrnto be the next best after the random forest in terms of overall performance metrics.rnHowever, the support vector machine classifier in both test modes exceedsrnthe acceptable detection delay in the classification process.