MONITORING THE DURATION OF TELEPHONE CONVERSATIONS USING ПAMMA DISTRIBUTION

Zubarev Kirill M., Makarichev Georgiy O.

This work is devoted to the construction of a model based on the gamma distribution, which allows identifying abnormal patterns in the data on the duration of telephone calls, which may indicate the presence of spam activity. The study used telephone call data that had undergone preliminary filtering and processing. A histogram of relative frequencies was constructed for the data, and point estimates of the gamma distribution parameters were obtained based on the method of moments. An interval estimate of the gamma distribution parameters was also made. At the final stage of the study, a method for identifying spam activity was proposed, based on checking the compliance of the sample of telephone calls with the gamma distribution with parameters within the constructed confidence intervals.

Keywords:gamma distribution, method of moments, confidence intervals, Pearson criterion, phone calls, spam.