Research On Recognition Method In Customer Purchase Intention Based On Intelligent Customer Service Learning

With the rapid development of artificial intelligence technology,intelligent customer service has gradually replaced manual customer service,which greatly improves work efficiency.The continuous advancement of deep learning technology has greatly promoted the application of chat robots in the commercial field.For a conversation system with a chat robot,deep learning technology can automatically learn feature expressions and generate replying strategies by using a large amount of data in a network conversation with small amount of manual operations.The strategy is to establish an intelligent human-computer interaction system driven by network big data between computers and humans,so that the chat robot can communicate with us normally and effectively,improving work efficiency and quality of life effectively.In recent years,the e-commerce industry has developed rapidly and using the relevant intelligent search technology to fully tap the business opportunities of online chat groups has become the current market trend.The application of intelligent chat robots to chat group purchase intention identification is a new application for smart chat robots.The so-called purchase intention identification refers to that the intelligent chat robot can fully explore the relevant product information expressed by the customer in the conversations and locate the purchase intention of the customer accurately,so as to conduct the customer's purchase intention effectively,and to recommend commodity effectively,and to improve the service greatly,and then increase users' satisfaction and profitability.Based on smart customer service,the method of customer purchase intention identification has been studied and a calculation model of consumer intention strength has been proposed.Firstly,to collect corpus data such as store merchandise information,reviews,and microblogging articles in the cosmetics field and to segment words in the corpus by using the segmentation technology of the Chinese Academy of Sciences.The custom dictionary involving commodities,beauty makeup,skin care,hairdressing,and other aspects has been added so as to improve the segementation accuracy.Then,according to the characteristics of people's daily communication language and thestatistics of the communication characteristics in the corpus,a stop word list has been established and the denoised corpus data shall conduct the word vector training by using the word2 vec training tool.The next step is to extract the keywords of the user's chat information,the realization method is: according to the corresponding chatting time point,to use the kmeans clustering algorithm which is based on the word2 vec to generate chat subject word records based on time series;according to Ebbinghaus forgetting curve to structure the forgetting function,to quantify the user's purchase intention intensity in multiple stages,and conduct experimental analysis;Finally,to compare the experimental results with the sentiment analysis results.The comparison results show that the consumption intention strength calculation model in this paper can clearly and intuitively lock the user's purchase intention,provide more accurate data support for the smart customer service,improve the customer service effect,and then increase the sales conversion rate.The research in this paper has valuable references for the development of consumer purchase intention identification technology in the field of smart customer service.