Real-time identification method for call result

Abstract

The invention provides a real-time identification method for a call result. The method comprises the following steps: a call center initiating a call; acquiring a voice stream containing a ring-back tone when the call is not answered; processing the voice stream to extract an audio fingerprint of the voice stream; searching a sample audio fingerprint matched with the audio fingerprint; and determining a recognition result corresponding to the voice stream according to the sample audio fingerprint. In addition, the invention also provides a corresponding computer medium. By implementation of the method, the requirement of real-time identification can be met while the identification accuracy is ensured, so that the call-out efficiency of a call center is improved.

Description

technical field [0001] The invention relates to the field of service design of a call center, in particular to a method for real-time identification of call results. Background technique [0002] In the current application scenario where the call center makes an outbound call, it is usually necessary to judge the connection status of the target number for the outbound call. If the outbound call is not connected, it is more desirable to clarify the reason for the non-connection of the target number. In order to facilitate the call center to further formulate a call strategy for the target number. [0003] In the prior art, there are usually two methods for the call center to determine the reason for the failure of the outbound call not connected: method 1, to judge according to the result code returned by the operator; method 2, the call center to provide a suitable calculation The resource extracts the features of the recorded call voice, and further judges its specific con...

AI Technical Summary

Problems solved by technology

For method 1, since there is no uniform standard for the result code of the operator, and the operator will upgrade or adjust the result code in stages, and the corresponding logical matching upgrade of the call center is delayed, resulting in no connection The reason is that the accuracy of the recognition results is difficult to reach the expected value; for the second method, the computing resources occupied by speech recognition are relatively large, and in the case of limited computing resources, it cannot meet the real-time requirements for immediately returning the recognition results when making an outbound call. Due to the limitation of computing resources, the existing speech recognition schemes can only recognize and classify fixed-length speech segments, and cannot achieve better recognition results for each different stateful speech stream, especially Unable to recognize user-defined voice clips such as rejecting CRBT

Whether it is method 1 or method 2, the inability to accurately identify the reason why the outbound call is not connected in real time will affect the further improvement of the call center's outbound call efficiency

Images