Implementation And Application Of Human-Computer Dialog System In The Field Of Mobilecommunications

With the telecommunications operators user scale continues to expand, through the telephone complaints, consulting and business deal with service number are also constantly increased, this will be on phone voice access technology put forward higher request, man-machine conversation automation system should potential is born. Now the operators are already set up enterprise call center to be automatic response system, this system can perform simple play the clues, carries on the key function menu option. But such simple function is far from enough, the enterprise decision makers hope to build a more functions more comprehensive system, can effectively solve the user’s question, intelligent judgment and handling problem. So we how to develop automation man-machine conversation system (HCDS) makes a deep exploration and research, and obtain some results, the theory for the future will establish a new intelligent man-machine conversation system lay the foundation.In this paper, we expounded how to develop automation man-machine conversation system (HCDS), this system is a kind of application software to enterprise self-built automatic answering customer service telephone system as the foundation, the design and realize the intelligent man-machine conversation function. Thesis mainly work is divided into the following three parts:First of all, we explore the traditional service automatic answering telephone system in the limitations of the telephone customer service, this paper analyzes the intelligent man-machine conversation system in the enterprise the important significance of call center, this paper expounds the intelligent man-machine conversation system realization structure and core technology.Secondly, the thesis on the man-machine dialogue system for the overall frame design. In the software development process, we put forward by using a kind of probability method makes every time we can predict the man-machine dialogue user satisfaction. To do this, in our system will be used to be automatic extraction dialogue function, it is a kind of fast steady machine automatic learning model, we also use the hidden markov model (HMM) to establish what we expect to realize system. In order to evaluate the performance of the system, we use two public corpus:2010and2011DARPA corpus.Finally, we use10times layered cross experiment to verify the system evaluation. Our experimental results show that the system can applied to real life, in the future for the intelligent equipment development carries on the service.At present, based on the speech recognition system has been produced, and successfully applied to the chongqing mobile customer service center outbound platform. After the test, this system has a good concurrency and stability, and at the same time, through the number of effective screening and screening, significantly reduce the position waiting duration, significantly reduce the position waiting time and improve the dialing efficiency, and then reduce the outbound cost.As call centers markets mature, based on man-machine conversation recognition technology will improve the efficiency of the call center and increase the profit of enterprise has the extremely vital significance, its application prospect will be very broad. But in this field, there are still many problems worth studying. In the future work, we will be more inclined to based on the development of the new prediction technology of man-machine conversation automatic identification system.