User Comment Analysis Of Household Water Purifier Based On Text Mining

With the rapid development of Internet and the transformation of logistics technology,e-commerce online shopping has become an efficient channel for consumers to purchase household appliances.People’s attention to drinking water sanitation makes household water purifier,an efficient and convenient small household appliance,become the first choice of more and more Chinese families.The huge trading volume of e-commerce platform has brought a large number of product reviews.User reviews are consumers’ intuitive feedback on the purchased products.Through user reviews,potential consumers can truly understand the product characteristics and purchase products that meet their own needs.Enterprises and merchants can improve their own products and services,so as to win in the market competition.However,there are many problems in e-commerce reviews,such as disorderly structure and large quantity,which are not conducive to the extraction of effective information.Therefore,the efficient analysis of text reviews is of great practical significance to consumers’ purchase decisions,enterprises’ production and sales,and the operation and management of e-commerce platform.In this paper,more than 80000 user reviews of A.O.Smith,Qinyuan and Xiaomi household water purifiers in Jingdong shopping mall are taken as the object of study.The following research is carried out: first,emotional phrases are constructed to merge and expand the existing emotional dictionary,emotional analysis is carried out on user reviews of three brands of household water purifiers,emotional scores of reviews are calculated and classified.Secondly,LDA theme model is used to extract the theme,characteristic words and corresponding probability of positive and negative comments of each brand,and to summarize the advantages and disadvantages of each brand’s domestic water purifier.Thirdly,taking the quantified LDA comment topic as the input variable and the emotional intensity of the classified comment as the output variable,a BP neural network model is established to mine the main factors that affect the good and bad user satisfaction.Finally,this paper summarizes the research results and puts forward relevant suggestions.The results show that the average accuracy of emotional classification of household water purifiers based on emotional words combination and extended emotional dictionary is 81%,and the classification effect is good;after extracting the positive and negative comments of various brands of household water purifiers,A.O.Smith’s main concern factors are brand and water quality through BP neural network mining out a.o.Smith’s main concern factors are brand and water quality,while the main concern factors of poor users are quality Noise and sales promotion;Qinyuan praised the main factors concerned by users are distribution and water quality.Users with poor reviews mainly focus on after-sales and operability;users with good reviews mainly focus on cost performance,intelligence and appearance,while users with poor reviews mainly focus on quality and quotation.From the perspective of brand merchants,A.O.Smith should increase the research on the noise reduction function of products,and improve the sales promotion methods to achieve honest sales and reduce the brand premium;Qinyuan needs to improve the operation design of products to strive for more convenient and humanized use of products.At the same time,increase the research and development of filter elements to reduce the wastewater ratio of domestic water purifier products;millet needs to develop more high-end product lines to meet the needs of different levels of water purification.At the same time,we should also increase the overall safety inspection of products before they leave the factory,formulate reasonable and stable sales prices,and improve quotation services.From the perspective of e-commerce platform,JD needs to expand and improve the logistics and transportation area,and improve the logistics timeliness in remote areas.Strengthen the logistics packaging,improve the after-sales service attitude,and put an end to the secondary sales of return machines.