Research On Real-time Sentiment Analysis And Key Phrase Extraction Technology In WeChat-based Customer Service System

As the main way to interact with customers,customer service system plays an important role in understanding customer needs,solving customer problems,mastering market trends,and enhancing corporate image.Traditional customer service systems based on instant messaging tools or online automatic question answering systems has many flaws.Such as waste of human resources,difficult to share resources,low accuracy and so on.Research and development of a new generation of universal,versatile,personalized and intelligent customer service system of great significance whether for users or enterprises.With the development of mobile Internet and the popularity of smart phones,WeChat plays an increasingly important role in people’s lives and has a broad user base.So WeChat in customer interaction and mark eting advocacy has great value.In this paper,a new intelligent customer service system is designed and developed based on the characteristics of customer service system and WeChat,and real-time emotion analysis and user key phrase extraction technology are developed under the support of this project and details as follows.First a real-time emotion analysis method based on short-term and short-term memory network is proposed for intant messaging chat text.The ability to automatically analyze user opinions and emotions is called real-time emotional analysis.Real-time emotion is an important part of intelligent customer service system as a powerful resource.Through the Long Short-Term Memory(LSTM)model,the neighborhood position information is integrated,which alleviates the problem of short text feature sparseness to a certain extent.The time series problem is solved by the LSTM model,and the effect of the emotion analysis can be achieved in real time.The effectiveness of the method is verified by simulation experiments.Second,a text key phrase extraction method for user chat records is proposed.Mining key phrases and other key features of the customer can be accurate marketing and other aspects of reserve knowledge.Based on the TextRank algorithm,word candidates are extracted from candidate words based on word theme similarity degree and word topic influence,and then stronger phrases are generated according to candidate keywords.The experimental results show that this method can extract key phrases covering flie main topic information of the text.Finally,we design and implement the customer service system based on weChat.Real-time emotional analysis and user key phrase extraction technology applied to the actual task of the system to help our customer service system to the intelligent,personalized and diversified direction a step forward.