The Design And Implementation Of Voice Group Call Analysis Platform

A phenomenon that someone calls a large quantity of user phone number in a short period of time is named as the voice group calling. It is a calling method for marketing network platform based on VOIP. Many criminals use this method to do the malicious voice group call. They usually call a quantity of users with a short ring, to make them callback.Then, many user will be deceived by the malicious voice group call. This calling situation is named harassing voice group call. The harassing voice group call is very common, most of them are related to marketing fraud callings. To a greatly extent, this voice group call has threatened the normal safety of user's property, and reduced the user's quality of life, and largely occupied the network resources of operators. In terms of the rampant situation of the harassing voice group call with telephone fraud, there are no mature effective solutions currently. To solve this problem, a Telecom operator collects user's reports and manually dialects back to check whether the number is a harassing voice group call. However, this solution costs too much manual labour and with low efficiency. On the other hand,as far as the solution of harassment telephoneare concerned, the China mobile of Hunan province currently has its own analysis of strategy. But according to the voice group call, the analysis strategy of Hunan China Mobile can't produce satisfying result. Because Hunan China mobile isn't devoted to the harassing voice group call, the recognition of harassing voice group call is not approving, with a poor accuracy, and it is hard to largely cover all harassing voice group call.So in this paper, based on the present network environment of Hunan China mobile, by the design of harassing voice group call's quota and the model analysis and the rules extracted from the model, we design and implement an analysis platform to solve the harassing voice group call. The entire platform has successfully been working in the laboratory and Hunan China Mobile, the current network environment. At the same time, we calculate the platform from accuracy and coverage to evaluate the result of the platform. The result is with 60 percent on accuracy and 49.3 percent on coverage. The whole system can identify the harassing voice group call accurately, automately and efficiently. As a more significant result, the platform will greatly reduce the cost of manual auditing. In this thesis, a solution of harassing voice group call using a modeling analysis approach is put forward, from the technical level and with the big data support. By this solution, the operator can keep the user's property safe, and improve the quality of service, and reduce the waste of network resources.