| Customer service system is an important component of manufacturing enterprises.Especially in complex product manufacturing enterprises such as automobile enterprises,knowledge service is one of the core competitiveness of enterprises.Question answering system(QA)is a kind of knowledge service technology which is very suitable for enterprises to serve customers.Users can get the information they want quickly and concisely through the QA,which improves the efficiency of information acquisition.In view of the complexity of the QA and the particularity of the automobile industry,this dissertation studies the architecture design of the QA for automobile manufacturing enterprises,and puts forward corresponding improved algorithms for the key technologies in the process of the QA.Firstly,based on the analysis of the service characteristics of the automobile industry,this dissertation designs the framework of the QA based on the community question answering text,which makes the general QA more suitable for the needs of users in the automobile field and helps automobile enterprises to actively acquire the hot issues concerned by users.Secondly,in view of the complex and serious spoken questions faced by the community-based automobile QA,this dissertation studies an improved automobile question text classification algorithm based on deep learning.In this dissertation,the application of deep learning model in text classification and attention mechanism of question words is studied.Finally,this dissertation studies the answer selection algorithm for joint modeling of automotive question-and-answer text,which provides theoretical support for answer selection and answer ranking in QA.In order to further illustrate the effectiveness of the two improved algorithms proposed in this dissertation,experiments are used to verify and analyze the results.The experiments show that the algorithm proposed in this dissertation achieves the expected research results. |
PDF Full Text Request