Design And Implementation Of Intelligent Auxiliary Processing System For Personal Calls Under Meeting Scenarios

In the research of current application in the field of smart office,"people-centered" service concept has been a broad consensus. By integrating context-aware technology and data mining technology, this paper design and implement an intelligent auxiliary processing system for personal calls under meeting scenarios. The system can take different intelligent auxiliary processing measures for different callers under different meeting scenarios, which allows participants neither missing important calls, nor being affected by non-critical calls during the meeting, so that providing a good environment for the conference participants. Contents of this paper are as follows:(1) Research and implement Bayesian network instruction and inference under the evidence completion strategy, and complete participants state uncertainty reasoning under meeting scenarios. For Bayesian network model, this paper compares different structure learning and parameter learning algorithm combined with the specific experiment, and on this basis complete the construction of Bayesian network; For the problem of low reasoning accuracy in the Bayesian network inference due to lack of evidence, this paper presents an evidence completion strategy based on historical context, and experiments show that the strategy can improve the accuracy of participants state uncertainty reasoning while lacking of enough evidence.(2) Research and implement a method to obtain contacts importance of time distinguished based on communication of history data. This method firstly calculate the communication closeness and communication time distribution between the user and the contacts, and then get the contacts importance of time distinguished, which provide a scientific basis for automatic caller information processing for users under special circumstances. Experiments on public data sets verified the reasonable of the method proposed in this paper.(3) Design and implement dynamic adjustment and update mechanism for auxiliary processing system for personal calls under meeting scenarios. To solve the problem of too rigid and single of the processing measures, this paper design and implement dynamic adjustment and update mechanism for the auxiliary processing system. Specifically, add the dynamic update mechanism during the participants state reasoning by using Bayesian network, add the dynamic adjustment mechanism based on times of calls during the calculation of the contacts importance, add the dynamic adjustment mechanism for processing measures during the auxiliary processing. Experiments show that all these mechanisms makes the measures for personal calls given by the auxiliary processing system more reasonable.